Skip to main content

Full text of "Eclipses Of The Sun"

See other formats

< c ^ 

W > DO ^ 

u<OU_1 64026 >[g 

CD "" 

^ D 


From the painting by Howard Russell Butler, N. A. Shows the edge of the reappearing sun. 





PROFESSOR F"TTf l To*fr6" i ?TY AT 





This edition comprises all the material contained in the third edition, 
as well as a new chapter and certain additions to the text. 
Chapters in which no out-of-date scientific statements occur 
have been reprinted without change from the third edition. 

COPYRIGHT 1923, 1924, 1932, 1935 

First edition, 1923 

Second edition, revised and enlarged, 1924 
Third edition, revised and enlarged, 1932 
Fourth edition, revised and enlarged, 1935 






THIS book owes its inception to the keen scientific 
interest of Mr. Edward D. Adams. The magnifi- 
cent painting of the corona by Howard Russell 
Butler, which now finds an honored place in the American 
Museum of Natural History, New York, was originally 
planned that it might provide a frontispiece illustration. 

Fortunately for the author, he had the hearty coopera- 
tion of many of his friends in the preparation of the book. 
Professor E. E. Barnard, an intimate friend for twenty-five 
years, furnished all of the photographs illustrating the 
eclipse expedition in 1901 to Sumatra. The last work 
undertaken by him, a few days before his lamented death, 
was to dictate the captions to accompany each of the photo- 
graphs. Another life-long friend, Professor Edwin B. Frost, 
director of the Yerkes Observatory, was most generous in 
his assistance. As editor of the Astrophysical Journal, he was 
instrumental in providing the blocks for many of the illustra- 
tions. Professor George B. Pegram, chairman of the Adams 
Fellowship Committee of Columbia University, showed con- 
tinued interest during the progress of the work. Dr. W. W. 
Campbell, director of the Lick Observatory, kindly furnished 
information in advance of publication regarding the Einstein 
problem and provided photographs of the 1922 eclipse and 
with great generosity supplied many additional illustrations 
from the incomparable Lick series of photographs of the solar 
corona, while Professor H. Deslandres kindly sent for repro- 
duction many photographs taken with the spectro-heliograph 
at Meudon. 

The author wishes to express to these scientists his heart- 
felt appreciation for their assistance so freely given. 

The second edition was prepared while the author was 
in residence for a brief stay at the Mount Wilson Observa- 

In the eight years which separated the second and third edi- 


tions, scientists had a most thrilling time in their investiga- 
tions of the chemical atom. It has been said with truth that 
discoveries have followed each other in such rapid succession 
that before the printer's ink was dry, a new and, at times, 
startling theory was superseded by another theory still newer. 

As the problems of solar physics have their fundamental 
basis in the structure of the atom, an attempt has been made 
to incorporate in this book the recent important researches 
dealing with atomic physics. 

The fourth edition, prepared while in residence again at 
Mount Wilson Observatory, adds to the third edition a chap- 
ter giving the results obtained from the 1932 and 1934 
eclipses. Minor changes have been made in Chapter IV deal- 
ing with future eclipses. 


May 13, 1935 


INTRODUCTION xiii-xviii 
























XIII. ECLIPSES SINCE 1923 203-225 


















XX. THE CORONA 360-386 









XXV. THE ECLIPSES OF 1932 AND 1934 483-511 


INDEX 513-520 


TOTAL ECLIPSE OF 1923, IN COLOR (Butler) Frontispiece 


CORONA AT 1930 ECLIPSE (Marriott) 9 

INNER CORONA, 1930 (Marriott) 12 

THE CORONA, 1929 (Swarthmore) 13 







TOTAL ECLIPSE OF JUNE 8, 1918, IN COLOR (Butler) 60 





THE SUN'S SURFACE, SEPTEMBER 1883 (Janssen) 116 

SUN-SPOT GROUP (Meudon Observatory) 117 






CORONA OF MAY 28, 1900 (Barnard and Ritchey) 164 








THE " HELIOSAURUS "AT 1918 ECLIPSE (Barnard) 196 



DISTURBED REGION IN 1923 CORONA (Swarthmore) 209 

CORONA OF 1923 (Swarthmore) 212 

THE CORONA OF JANUARY 24, 1925 (Slocum) 213 

THE CORONA OF JANUARY 24, 1925, IN COLOR (Butler) 224 










TOTAL ECLIPSE OF MAY 29, 1919 273 







BIPOLAR SUN-SPOT (Mt. Wilson) 333 

SOLAR VORTICES (Mt. Wilson) 340 



THE TOTAL ECLIPSE OF 1918 (Lick) 349 


TOTAL ECLIPSE OF 1922 (Lick) 365 







INNER CORONA IN 1926 (Swarthmorc) 400 




CORONA IN 1922 (Lick) 448 

FINE STRUCTURE IN CORONA, 1929 (Swarthmore) 449 



CORONA IN 1932 (Lick) 476 




FLASH SPECTRUM IN 1932 (Greenwich) 492 

FLASH SPECTRUM IN 1932 (Lick) 493 

CORONA IN 1934 (Satome) 500 



SCIENCE seems bent on increasing man's estimate of 
the span of time since creation began. It is now 
confidently believed that the human race has in- 
habited this terrestrial ball for at least five hundred thousand 
years, while Mother Earth herself must have been in exist- 
ence for a thousand million years. On the assumption that 
mass is electrical and that all kinds of energy possess mass, 
we have found that the sun's heat can not be maintained 
for a space of time exceeding fifteen millions of millions of 
years, provided all its energy, including that of the atom, 
could be liberated. Surely such an interval is sufficiently 
great to satisfy even the most exacting! And yet what 
changes have taken place in our mode of life within the 
memory of man! 

Today, as never before, our daily life follows its course 
surrounded by the wonders of science. The age of steam has 
brought the powerful locomotive drawing its fast and com- 
fortable passenger train on its long journey across the con- 
tinent, has constructed the gigantic ocean steamship where- 
by Europe is brought within four days of the shores of 
America, has been, indeed, the cause of the remarkable de- 
velopment experienced in all quarters of the globe. The 
internal combustion engine has largely revolutionized modes 
3f transportation on the land, on the surface of and under 
the sea and in the air, so that the extravagantly fanciful 
writings of Jules Verne are common-place, every-day reali- 
ties. The age of electricity has wrought the greatest 
:hanges of all in our daily life bringing as it has stupendous 
transformations in illumination vastly different from the 
Feeble tallow dip; it has wrought the powerful motor; it 
las made possible the electric telegraph and telephone and 
:he still more wonderful wireless, with the result that dis- 
;ance is eliminated and there is served fresh each morning 



with our coffee the very latest news garnered from every 
quarter of the globe. Through the agency of the cinema 
and the wireless telephone the inhabitant of a backwoods 
village can gaze upon great actors, can listen to the greatest 
of singers and the best of symphony orchestras quite as well 
as if he lived in London, Paris or New York. 

The bridge separating pure from applied science has been 
so much shortened that the gap has practically disappeared. 
The refined researches of the physicist or chemist today 
may be applied in the arts and sciences of tomorrow. In- 
vestigations in Hertzian waves in the laboratory have made 
the wireless telephone possible, the discovery of Mme. Curie 
and the remarkable work of Rutherford and others on the 
penetrating powers of the rays of radium have given to the 
cancer specialist the means of alleviating the horrors of this 
dread disease ; X-rays are now in every-day use by the doc- 
tor, the surgeon and the dentist. And in the quiet rest of 
the laboratory, the chemist has experimented with his vials 
and test tubes and has made noxious and poisonous gases 
which, liberated amidst the din and roar of the battlefield, 
have caused the destruction of thousands of lives. 

It is less than forty years since the discovery of X-rays 
and radium. The marvelous researches connected with 
radioactivity have shown the amazing degree of refinement 
necessary in the work of the scientist. New forces, hitherto 
unknown, have been discovered, while recent researches 
display the amazing fact that each chemical atom takes 
its place in Nature as a miniature solar system. Modern 
investigations have revealed that one chemical element may 
be transformed into another, thus realizing the age-long 
dream of the alchemist, the transmutation of one metal into 
another. It has been proved that several separate series of 
radioactive transformations have as their end-product the 
metal lead but fortunately for the economic future of the 
world, the realization of the alchemist's vision of changing 
base lead into glittering gold seems as far distant as ever. 
Helium is a product of almost every radioactive transforma- 
tion and it has thus become a household word with every 
trained physicist. And yet helium was discovered a re- 


suit of solar eclipse observations a bare half-century ago 
and was isolated from pitchblende only in 1895! Within 
two decades the demands of the Great War brought to the 
attention of scientists the urgent necessity of obviating the 
disastrous explosions that have taken place with hydrogen- 
filled balloons with the consequent result that helium was 
produced in such large quantities that balloons were filled 
with this inert gas. 

But amongst all the wonders of all the wonderful sciences 
there is no science which deals with such a gorgeous spec- 
tacle as is exhibited by the queen of the sciences, astronomy, 
at the moment when the earth is gradually shrouded in dark- 
ness and when around the smiling orb of day there appears 
the matchless crown of glory, the so-called corona. Nor can 
any science duplicate the wonderful precision shown by the 
work of the astronomer in his capacity to predict hundreds 
of years in advance the exact hour and minute at which an 
eclipse will take place and the locality on the earth's surface 
where such an eclipse will be visible. 

The great progress of science in the last fifty years is 
nowhere better illustrated than in the attitude of astrono- 
mers towards observations at the time of a total eclipse of 
the sun. Until about the middle of the nineteenth century 
little interest was taken in the subject although information 
regarding the sun was comparatively scant. The eclipse 
was observed only if perchance its track happened to cross 
the home of the observer; the only observations of value 
being the exact times of contact of the limbs of the sun and 
moon, taken for the purpose of perfecting the lunar tables. 
The beautiful corona was watched with awe and admiration, 
a few sketches were made of its form, but these were 
done with such indifferent skill that they added but little 
to the information available at the time. Practically no 
expeditions were equipped and sent away from home. How 
different it is in the twentieth century! In the year* 1901, 
the United States Naval Observatory financed and sent an 
expedition as far from home as it could possibly go half 
way round the world and all for the purpose of making 
observations during a few short minutes of time. At the 


recurrence of the eclipse in 1919, the British observers in 
Brazil and West Africa startled the thinking world by veri- 
fication of the Einstein problem. In September, 1922, the 
British astronomers had spent more than six months on 
Christmas Island in the Indian Ocean only to have their work 
come to naught because of clouds at the critical moment, 
while American and Canadian scientists had gone to North- 
west Australia in quest of information about the relativity 
shift. On isolated " Tin-Can Island " in the South Pacific, 
important discoveries were made in 1930. 

To make such expensive and time-consuming expeditions 
worth while there must be problems connected with the sun 
whose solution is of vital importance to the astronomer. In 
this book the author will endeavor to state some of these 
problems and the methods devised for their solution. At 
the same time it will be necessary to refute one of the 
" twice-told tales " (See Science during 1921) that the moon 
is a more important body than the sun since the moon gives 
us light at night when it is dark and we need its light, 
whereas the sun shines during the daytime when it is bright 
and we could possibly get along very well without it! The 
light that lightens the world, the heat that gives our bodies 
comfort, the wind that cools our heated brows and wafts the 
sailing ships across the ocean, the coal that warms and illumi- 
nates our homes and generates steam and electricity to carry 
on the world's mighty commerce, the rain that descends 
from heaven and waters and fertilizes the soil and causes 
the flow of our mighty rivers; all these and many other 
benefits come from the sun. Without the sun there would 
be no grass, no flowers, no wheat or corn or vegetable life 
of any kind, and without the sun there would be no animal 
life, no man upon the earth. If the sun were blotted out 
for the space of one short month, there would not be one of 
us left alive to tell the tale; we should all be frozen to death! 
It is well then that we should endeavor to learn as much as 
possible of the giant luminary in which are centered our light 
and heat and our very life, even when the quest for knowl- 
edge carries one as far from home as does eclipse work. 

The life of such an observer might be likened to that of 


a hunter after big game. Many months and even years are 
spent in quietly investigating the problems, a costly equip- 
ment is accumulated and each piece of delicate apparatus 
is carefully tested at home to see that it will properly per- 
form its designated functions far afield, a long journey is 
often necessary, frequently of thousands of miles, by rail 
and sea. Arrived at the destination, instruments, cameras 
and spectroscopes are erected and most carefully adjusted, 
and after six or eight weeks of preparation in the field the 
eventful day approaches. Each and everyone of the party 
drills constantly so that the task allotted to him may be well 
and carefully done, so that the photographic slides may be 
drawn and each camera lens may be opened at the appro- 
priate instant. Success lies in seeing that every one of a 
thousand possible chances of failure are obviated. At a cer- 
tain hour, minute and second, the " zero-hour," operations 
are due to begin. But alas! there may be no " game," the 
eclipse may be eclipsed by clouds, and the long months of 
preparation may be of no avail since it will not be possible 
to try again on the morrow when the clouds have rolled 

The author has traveled ninety thousand miles to witness 
nine total eclipses of the sun. The total accumulation of time 
afforded him for scientific observations during these nine 
eclipses has been approximately eighteen minutes. 



"In old Cathay, in far Cathay, 

Before the western world began, 
They saw the moving font of day 

Eclipsed, as by a shadowy fan; 
They stood upon their Chinese wall. 

They saw his fire to ashes fade, 
And felt the deeper slumber fall 

On domes of pearl and towers of jade." NOYES. 

THE earliest recorded eclipse of the sun is one which 
happened more than four thousand years ago, 
an account of which is given in the ancient Chinese 
classic Shu Ching. According to the competent authority 
Oppolzer, this took place on October 22, 2137 B.C., about 
1400 years more remote than that recorded by any other 
nation. This eclipse is celebrated not only -for its great 
antiquity, but also for the dire fate of the two royal astron- 
omers, Hsi and Ho, who instead of staying in the sober 
paths of science went and got beastly drunk, with the re- 
sult that they were taken unawares and were unprepared 
to perform their customary rites of shooting arrows, beat- 
ing drums, etc., for the purpose of delivering the sun from 
the monster that was devouring it. To show his great dis- 
pleasure, not so much for failing to predict the eclipse, but 
on account of the intense confusion that prevailed, Chung 
K'ang, the fourth emperor of the Hsai dynasty, ordered 
that they should be punished and that their heads should 
be chopped off. And with this tragic warning in view, there 
is no record from that day to this that an astronomer has 
ever dared to follow in the steps of the unfortunates, Hsi 
and Ho ; and been drunk at the time of an eclipse. 


This eclipse is of such great importance that it will be 
well to give it more than a passing glance. The early 
eclipses, both of sun and of moon, have been of such great 
interest to astronomers and scholars that a very large num- 
ber of investigations have been devoted to the determina- 
tion of as exact dates as possible for these phenomena. 
Eclipses can take place only when the centers of sun, earth 
and moon are approximately in a straight line. These cir- 
cumstances can be predicted with accuracy if we know with 
a sufficient degree of precision the motions of the earth 
about the sun, and of the moon about the earth. The times 
of eclipses thus give to the astronomer the means of ac- 
curately testing his calculations regarding the motions of 
both these heavenly bodies. The movements of the earth 
render no difficulties, but with the earth's child, the moon, 
the matter is entirely different. It may be truly said that 
the motions of the moon have given the mathematical as- 
tronomer more work and worry than those of all the balance 
of the gigantic universe put together. The earlier the eclipse 
that can be checked up with accuracy the more reliably can 
the motion of the moon be verified, since any accelerations 
in this motion depend on the square of the elapsed time. 
Valuable as these early eclipses are to the astronomer, they 
are equally important to the historian or chronicler in fix- 
ing the dates of remote antiquity. The astronomical prob- 
lem has been well studied and is now regarded as a simple 
one. The crux of the whole question lies in the exactness 
and precision with which the event has been described by 
the author or historian. 

The Shu Ching, or Book of Historical Documents, is a 
collection of public speeches and proclamations beginning 
with the reign of the legendary Emperor Yao who lived in 
the twenty-fourth century B.C., and closing with the year 
B.C. 625. The book is not a historical, chronological narra- 
tive, nor indeed does such a book exist in the Chinese lan- 
guage. This ancient eclipse has been fully discussed by a 
number of eminent authorities. The most complete mono- 
graph is by Schlegel and Kiihnert, Die Shu-King-Finsterniss, 
Verhand. der Konin. Akad. van Wettenschappen, Letter- 


kunde 19, 5, 1890. Oppolzer communicated to Monats- 
berichte der Kbn. Preuss. Akad. der Wissens. zu Berlin, 166, 
1880. Briefer accounts may be found by S. M. Russell, 
Observatory, 18, 323, 1895; in Chambers' The Story of 
Eclipses, 65, 1900; and in the Halley Lecture delivered by 
Dr. J. K. Fotheringham, May 17, 1921. 

In the third century B.C., Shi Hwang- Yi, a great military 
genius, conquered the whole of the China of those days. 
In 221 B.C. he proclaimed himself the " First Emperor," and 
he decided, on the advice of his prime minister, that every 
form of human progress including literature should begin 
with his reign. He therefore ordered the destruction of all 
books except those belonging to the three utilitarian 
branches of knowledge, agriculture, divination and medicine. 
To see that his orders were properly carried out, he per- 
sonally examined each day a hundred pounds of books so 
that he might decide which were useless. Four hundred 
and sixty scholars were put to death for disobeying the royal 
commands, and others were banished for life. All copies 
of the Shu Ching seem to have perished except one incom- 
plete copy later recovered from a receptacle where it had 
been walled up by a devoted scholar. The book in which the 
eclipse is mentioned is not included in the authentic copy. 
It was added later, probably to conform to a table of con- 
tents that scholars believe may have been included in the 
original volume, and which may possibly have been written 
by the great Confucius. The preface to this book is as 

" Hsi and Ho, sunk in wine and excess, neglected the 
ordering of the seasons, and allowed the days to get into 
confusion. The prince of Yin went to punish them. De- 
scription of this there was made, The Punitive Expedition 
of Yin." 

It should be noted in the above that there is no mention 
of an eclipse. Tso, a scholar and commentator of the fifth 
century before Christ wrote concerning the lost work, The 
Punitive Expedition of Yin in the following words, " The 
Sun and Moon did not meet harmoniously in Fang. The 
blind beat their drums; the inferior officers galloped and 


the common people ran about." There is further added, 
" That is said of the first day of this month; it was in the 
fourth month of Hsia, which is called the first month of 

In the Annals of the Bamboo Books, a work of the early 
third century before Christ, the reference to Hsi and Ho is 
as follows: " In the fifth year of Chung K'ang, in the 
autumn, in the ninth month, on the first day of the month, 
there was an eclipse of the Sun, when he ordered the prince 
of Yin to lead the imperial forces to punish Hsi and Ho." 

The text does not say that the expedition was for the 
purpose of punishing the astronomers for failing to predict 
the eclipse. The scholar who restored the missing text used 
such flowery language that it is difficult to obtain the exact 
meaning, although it seems to be implied that the eclipse 
was the cause of the punitive expedition. 

There are many uncertainties that prevent an accurate 
interpretation of this eclipse, chief among which may be 
noted: (i) The month is entirely unknown, whether it is 
the " first day of the last month of autumn," or, " the first 
day of the first month of summer." (2) The meaning of 
the word Fang is doubtful, whether it means " the order 
of the constellations," which seems to be the better opinion, 
or whether Fang is used in the more restricted sense as the 
name of a small constellation including /3, 6, TT and p 
Scorpii. (3) The Emperor's name is very uncertain as it is 
not given except in the Bamboo Books. Little is actually 
known of Chinese chronology further back than B.C. 841. 
We do know the names of the emperors in succession, but 
whether they reigned five years or fifty is unknown, and 
consequently all that it is possible to say about the emperor 
Chung K'ang is that he must have been on the throne a 
century or two earlier or later than the year 2000 B.C. 
(According to the Encyclopaedia Britannica the first of the 
kings mentioned in the Shu Ching reigned from 2357 to 
2255 B - c -> an interval of no less than 102 years!) 

One interpretation of the above uncertainties is probably 
about as good as another. One date fixed by the Chinese 
astronomers of about the eighth century A.D., was the year 


B.C. 2155. But in consequence of the fact that the eclipse 
of that year was invisible in China, there has arisen the 
well-known ditty: 

" Here lie the bones of Ho and Hi, 
Whose fate though sad was risible, 
Being hanged because they could not spy 
The eclipse which was invisible." 

Taking " the last month of Autumn," and Fang in the 
restricted meaning of the word that it referred to the con- 
stellation, Oppolzer fixes the date as given above, B.C. 2137, 
October 22. Other authorities equally competent give other 
dates. The only conclusion that can be reasonably drawn 
is that it is not possible to identify the eclipse with any 
approximation to certainty. Nothing is known of the fates 
of the astronomers Hsi and Ho. The regulations regarding 
eclipses as given in the Shu King read: " Being before the 
time, the astronomers are to be killed without respite; and 
being behind the time, they are to be slain without re- 
prieve." With such a fate in store, who would dare be an 

Many admirers of the Chinese, not being conversant with 
the true facts, have pointed out this eclipse as evidence of 
the great learning of the Chinese and their proficiency in 
astronomical knowledge twenty centuries before the be- 
ginning of the Christian era. Although not wishing to dim 
the halo of glory that has surrounded early China, honesty 
compels one to draw attention to the facts. For many 
centuries the Chinese were unable to predict the position of 
the sun accurately among the stars for determining the 
length of the year. In fact, they seem to have relied wholly 
on observation from year to year to settle their calendar. 
No conclusions or systematic deductions were made from 
their observations as is shown by the fact that their calendar 
was continually falling into confusion. 

To those who are interested in early chronological data to 
be fixed by this and other remote eclipses, reference should 
be directed also to Delambre, Histoire de I' Astronomic 
Ancienne, Paris 1817, and to Johnson's Historical and 
Future Eclipses, London, 1896. 



The history of the Chinese Empire is very closely asso- 
ciated with the name of Confucius, the immortal sage. 
Confucius is the Latinized form of K'ung tsze (meaning 
the philosopher K'ung, this being the family name). 
Through his veins flowed the best blood of China. His 
birth has an interesting romance connected with it. His 
father Heih had a family of nine daughters but only one 
son who was unfortunately a cripple. Realizing in his old 
age that the K'ung name would probably become extinct 
with him, he went to his neighbor of the Yen clan, and told 
of his plight, and asked in marriage one of Yen's three 
daughters. The youngest of the three became the mother 
of Confucius, 551 or 550 B.C., Heih then being over seventy 
years of age. The father died in the child's third year. 

In his twenty-second year Confucius began his life as a 
teacher, and he put into practice principles which every col- 
lege professor of the twentieth century would like to follow, 
namely, he would accept any pupil no matter how small the 
fee, but as soon as lack of capacity or diligence was mani- 
fested, the youth was sent away. " When I have pre- 
sented," he is reported to have said, " one corner of a subject 
and the pup\l cannot of himself make out the other three, 
I do not repeat my lesson." Would that these methods of 
education were prevalent today! His professed disciples 
numbered 3000, of whom 70 or 80 were " scholars of 
extraordinary ability." Most of his life was spent in the 
province of Lu, the modern Ho-nan and Shantung. 

Among the writings ascribed to Confucius are the follow- 

(1) A preface to the Shu Ching, already referred to. 
The preface is little more than a table of contents, and it is 
very doubtful if Confucius had anything to do with it. 

(2) He compiled, or edited the Shih Ching, or Book of 
Poems. Originally numbering about 3000, only 311 were 
retained by Confucius, of which we now possess 305. It is 
the most ancient book of poetry in the world. The latest 
of the poems has the date 585 B,C. 


(3) From his own hand comes the Ch'un CWin, or 
" Spring and Autumn/' which is best known as the Annals 
of Lu, his native state. This work has been preserved al- 
most complete. It is a brief summary of the chief events 
that took place in Lu during a period of 242 years from 
722 to 481 B.C. It is a model for a historical document. 
The facts are briefly itemized according to the seasons in 
which they fell. As an example, in the year B.C. 612 there 
are twelve entries, the fifth of which was recorded, " In the 
autumn, in the seventh month, there was a comet which 
entered Pei-ton (in Ursa Major)." 

In the Annals of Lu are records of no less than thirty- 
six eclipses of the sun, the first eclipse being observed 
February 22, 720 B.C., and the last on July 22, 495 B.C. 
The first of them is described as follows: " In the 58th 
year of the 32nd cycle in the $ist year of the Emperor 
King-Wang of the Chou Dynasty, the 3rd year of Yin- 
Kung, Prince of Lu, in the spring, the second moon, on the 
day called Ke-Sze, there was an eclipse of the Sun." 
These ancient eclipses have been carefully investigated and 
many papers concerning them have been published in the 
Monthly Notices of the Royal Astronomical Society. The 
conclusions are that it has been possible to identify no less 
than thirty-two of the thirty-six. Apparently four of the 
eclipses, those of April, 645 B.C., June, 592 B.C., September 
19, 552 B.C., and June 18, 549 B.C. are in error and did not 
take place. The record is a history of eclipses that were ac- 
tually observed to have taken place. The inference is 
readily drawn that the astronomers who recorded these er- 
roneous eclipses saw some curious phenomenon around the 
sun; and remembering the sad fate that befell the two 
negligent astronomers, Hsi and Ho, and not wishing to take 
any chances whatever regarding the safety of their heads, 
they recorded the phenomenon as an eclipse. 

In addition to the works from the hand of Confucius him- 
self we are fortunately in possession of the Tso Chuan, a so- 
called commentary. This is presumably by some one with 
the name of Tso who took the bare entries in the work 
by Confucius and enlarged upon each one to such an extent 


and with such remarkable genius and dramatic brilliancy 
that the Tso commentary reads more like a prose epic than 
an elaboration of a series of facts or annotations on the 
text of a literary work. By its means there is vividly por- 
trayed the intrigues, the alliances, the treacheries and the 
jealousies of the various states which made up feudal China; 
we can see with the clearness almost of a photograph, as- 
sassinations, battles, heroic deeds, brilliant rescues and the 
torments and horrors meted out by the conquerors to their 
unfortunate victims. The Annals of Lu make the chronicles 
of China in the sth, 6th and yth centuries before the birth 
of Christ as full of action and as attractive as those of 
France and England twenty centuries later. As a matter 
of fact there appeared to be more of literary culture in 
China in these early centuries B.C. than there was in Europe 
a bare five hundred years ago. 

The eclipse of February 22, 720 B.C. recorded in the 
Annals of Lu is not the first. Chinese eclipse whose date can 
be accurately fixed, for fifty-six years earlier there was an 
eclipse of the sun preceded by an eclipse of the moon. In 
the Shift Ching, or Book of Poetry referred to above, there 
is an account which runs somewhat as follows: " Tenth 
moon, her conjunction first day, sin-mao, sun he had eclipse, 
also it very bad." The record ends by adding, " That moon 
in eclipse is a thing only common. This sun in eclipse is a 
thing very bad." The eclipse of the sun referred to took 
place on September 6, 776 B.C., the eclipse of the moon 
on August 2 r. The solar eclipse was visible only as a partial 
eclipse in China while that of the moon was nine-tenths 
total. It was during the reign of the notorius Yu-wang. 
It is apparent that the eclipse of the sun was regarded by 
the bard as a special warning to the lascivious Emperor. 
The evidence of this eclipse is indisputable; it fixes* abso- 
lutely without controversy the first certain date in <the 
history of China. How far before this epoch the Chinese 
history stretches is a matter of personal judgment and in- 
dividual conjecture. As already pointed out, the records 
of early dates are so unreliable that one of the early kings 
is supposed to have reigned the impossible period of 102 


Photograph by Marriott with 63 -foot tower telescope. 


years. Egypt has her pyramids and ruined temples as 
evidence of a much earlier civilization, but there are no 
such relics in China. This Chinese eclipse antedates by 
thirteen years the celebrated Nineveh eclipse. 

Mr. John Williams, formerly Assistant Secretary of the 
Royal Astronomical Society, made a careful examination 
of the Chinese historical work called Tung-Keen-Kang- 
Muh, a record of one hundred and one volumes which con- 
tains a summary from earliest times to 1368 A.D. Informa- 
tion concerning these eclipses can be found in the Monthly 
Notices R. A. S. Fifty-six solar eclipses occurred between 
481 B.C. and the beginning of the Christian era, and nearly 
one hundred additional ones before the fourth century A.D. 


" In Babylon, in Babylon, 

They baked their tablets of the clay; 
And, year by year, inscribed thereon 

The dark eclipses of their day; 
They saw the moving finger write 

Its Mene, Mene, on their sun. 
A mightier shadow cloaks their light, 

And clay is clay in Babylon." NOYES. 

According to L. W. King, Chronicles concerning the early 
Babylonian Kings, 2, 76, 1907, it is recorded that " on the 
twenty-sixth day of the month Sivan, in the seventh year, 
the day was turned to night, and fire in the midst of 
heaven. . . ." Cowell * has concluded that this probably re- 
fers to a solar eclipse visible in Babylon on July 31, 1063 B - c - 
By carrying the Babylonian records backwards from this 
eclipse, a check is had on the acknowledged Kea-Tsze 
6o-day calendar of the Chinese from which comes the date 
of the eclipse of 2137 B.C. Unfortunately, there is no 
knowledge concerning the name of the king in the seventh 
year of whose reign the eclipse occurred. If this fact were 
known, the early Babylonian chronology would be fixed 
with fairly great certainty. It is possible to conjecture as 
Langdon has done that the name of this king was Nabu- 
shum-libur, but this name must be looked upon only as a 

1 Monthly Notices, R. A. S., 65, 861, 1905. 


suggestion. All that is known of exact dates in Egyptian 
and Babylonian chronology depends on the science of as- 
tronomy, on the occurrence of an eclipse of the sun in 763 
B.C., and on the records of the astronomer Ptolemy as given 
in his great work the Almagest. 

There are no more interesting portions of the world's 
early history than that connected with the lands contiguous 
to the Persian Gulf. In Egypt, the gigantic pyramids, the 
construction of which is justly regarded as one of the seven 
wonders of the world, and the ancient temples point to a 
civilization that was in a high state of development five 
thousand years ago. Across the narrow Red Sea, more in- 
teresting even than the land of the Pharaohs, is the country 
of the Holy Scriptures where lived Adam and Noah, the 
great King David, the holy prophets and the Lord Jesus 
Christ. Up to the middle of the nineteenth century little 
was known of ancient Babylonia and Assyria other than the 
account given in the Bible. The country which is nearly 
enclosed by the two great rivers the Tigris and Euphrates 
from Bagdad to the Persian Gulf is bounded on the north 
by Mesopotamia, on the east by the plain of Elam, on the 
south by the Persian Gulf and on the west by the Arabian 
Desert. To the north, in Mesopotamia, the country is more 
or less mountainous, while to the south the terrain is flat 
and marshy. In ancient times the plain was covered with a 
complicated network of canals. The remarkable fertility 
of the soil brought such great prosperity to these peoples 
that they aspired to be the masters of the world. According 
to Herodotus (i, 193), wheat commonly returned two 
hundred-fold to the sower, while Pliny (N. H. xviii, 17) 
states that it was cut twice each year. The native historian 
Berossus remarked that wheat, barley, palms, apples and 
many kinds of shelled fruit grew wild. The country was 
studded thickly with towns. But alas, the neglect of the 
canals has changed the face of the land so completely that 
the fertility that was once the wonder of the ancient world 
has departed, and in its place there is nothing but a barren 
and desolate waste, part of the year the land being a series 
of swamps and marshes. 


The ancient city of Babylon was on the east bank of the 
Euphrates about seventy miles south of Bagdad. The plain 
of Babylonia was called Edin, the " Eden " of Genesis ii. 
The chief city of Assyria, Assud, was on the west side of the 
Tigris, while the three other great cities of Nineveh, Calah, 
and Arbela were on the eastern bank of the Tigris. 

The excavations at Nineveh by Botta and Layard opened 
up a new world. Layard after indescribable difficulties laid 
bare the palace of Sennacherib (705680 B.C.) and that of 
Assurbanipal (668-626 B.C.), and in the palace of the latter 
was found a great library of tablets which finally numbered 
no less than thirty thousand. These objects when not too 
massive were transported to the Museum of the Louvre 
or to the British Museum. The success of these discoveries 
stirred scholars throughout the world, and the different 
nations vied with each other in the prosecution of the inter- 
esting search. The British and French continued their in- 
vestigations, and in 1886 German archaeologists began their 
work. Since 1888 expeditions have been organized by the 
University of Pennsylvania and since 1903 by the University 
of Chicago. 

The language in which the ancient history is preserved 
is the so-called Babylonian-Assyrian wedge-writing, the 
cuneiform language of signs. No such thing as an alphabet 
was known. As the $ sign means much to the American, 
while to the resident of the British Isles the synonym of 
money is conveyed by a different symbol, , so these primi- 
tive people expressed themselves in their tablets entirely by 
signs. A wedge was used to make an impression in clay, and 
the different combinations served to express a variety of 
ideas. The decipherment of these tablets was a difficult 
task. The most important progress was made through the 
work of an Englishman, Henry C. Rawlinson, who observed 
at Behistun a rock stretching up almost 1700 feet above the 
plain, and at a point about 350 feet above its base there was 
a large space carefully smoothed off. On this was found 
a mass of inscriptions distributed in columns of various 
lengths. After long years of work he succeeded in trans- 
cribing the whole of the record. Careful study and long 


and arduous toil revealed the fact that the writing was in 
three different languages, old Persian, Elamite and the 
Babylonian. The decipherment was excessively difficult due 
to the enormous number of signs employed. For simple 
names there were about 600 independent and distinct signs 
made by combinations of wedges numbering anywhere from 
two to thirty. But two to six of these signs might be com- 
pounded to express more complicated ideas, with the result 
that there are something like 20,000 different signs known 
to students of Assyriology. The number of inscriptions so 
far found number about one hundred and fifty thousand. 

The earliest mention of Babylon appears to be about 
3800 B.C. The first great king was Sargon of Akkad who 
lived about 2750 B.C. Even greater was Khammurabi the 
Amorite who flourished about 2100 B.C. Concerning this 
latter monarch, who lived over four thousand years ago, 
we have positive information in a group of fifty-five of his 
letters edited by L. W. King, and in a monument giving 
a code of splendid laws. From his letters we learn that the 
Babylonians kept their record of time entirely by the moon, 
a new month beginning with each new moon. Twelve lunar 
months made up the year. Since the synodic month is of 
twenty-nine and a half days in length, it became necessary 
to add in the reckoning of the year a thirteenth month from 
time to time in order that the calendar should not get astray. 

This method of measuring the length of the year, which is 
both inaccurate and inconvenient, was followed by the 
Hebrews and the Persians. Even in the enlightened age of 
the twentieth century the Jewish and Mohammedan cal- 
endars are still based on the old custom. An intercalary 
month being inserted in the reign of Khammurabi, he ac- 
cordingly sent out a circular letter to his governors advising 
them that the delinquent tax gatherers should not take ad- 
vantage of the change in the calendar and that the taxes 
must be paid without delay. Schools flourished even at 
this remote period, and tablets have been found which evi- 
dently were exercises of the children as they struggled to 
learn the language of signs. In fact a Babylonian proverb 

Photograph by Marriott with 63-foot tower telescope. 


g 6 

O cfl 
g u 


has been unearthed which reads, " He who shall excel in 
tablet-writing shall shine like the sun." 

The kingdom of Assyria begins about 3000 B.C. Assyria 
was a warlike nation, and lived by the sword. In the eighth 
century B.C., their conquests had been pushed as far west- 
ward as Damascus and to the Hebrew city of Samaria. 
While besieging this latter city, the king was deposed and a 
usurper called Sargon took the throne. Under him and his 
son Sennacherib, Assyria was at the zenith of its military 
power. Campaigns were carried out in Ionia and Greece 
and also southward as far as Egypt, which became tributary 
to Assyria. In Nineveh, Sennacherib constructed the largest 
and most magnificent palace the world had seen up to this 
time. In 689 B.C. he conquered the city of Babylon, and he 
caused this holy city to be completely destroyed and all 
of its buildings razed to the ground. After many years 
Babylon under Nabopolassar, revolted from the Assyrian 
yoke and under his son Nebuchadrezzar the Nebuchad- 
nezzar of the Bible construction on a great scale was 
carried out, and Babylon was rebuilt and became the wonder 
city of the ancient world. According to Herodotus, the 
walls were no less than 56 miles in circumference, the walls 
being 335 feet high and 85 feet wide. Each city had its 
great temple and its own god, the god of Babylon being 
Bel-Merodach. In the temple stood an enormous image of 
the god forty feet high, together with a table, a mercy seat 
and an altar, all constructed of gold. In this reign (604- 
561 B.C.), the rebuilt Babylon was at the height of her 
glory and the summit of Chaldean civilization had been 

This glimpse of the history of these early peoples of such 
great interest to all Christian countries is for the purpose of 
giving a setting for ascertaining what they had accomplished 
in art, literature and science particularly in the science of 
astronomy. Assyria differed much from Babylonia. The 
former state was an armed camp and held sway by the 
sword, probably even in those early times believing that 
might is right. On the other hand, the Babylonians were 
peace-loving people, a land of merchants and farmers, and 


withal deeply religious. Consequently, Assyria had little 
culture of its own and the little it did possess was borrowed 
or acquired from. Babylonia. 

One of the grandest epics of ancient Babylonia was the 
description of Creation which is mentioned here on account 
of its astronomical significance and also that it may be com- 
pared with the Biblical description. In the first book an 
account is given of the creation of the world out of the 
primeval deep and the birth of the gods of light. Then 
comes the story of the struggle between the gods of light 
and the powers of darkness, and the final victory of Mero- 
dach, who clove the dragon of chaos, Tiamat, asunder, 
forming the heaven out of one half of her body and the 
earth out of the other. Merodach next arranged the stars 
in order, along with the sun and moon, and gave them laws 
which they were never to transgress. After this the plants 
and animals were created, and finally man. Merodach here 
takes the place of Ea, who appears as the creator in the older 
legends, and is said to have fashioned man out of clay. 

Thus from very earliest times, the religion of the Baby- 
lonians was inspired by the belief that the heavenly bodies 
were subject to law and order and that they ruled the des- 
tinies of man. It was evident even to the simple mind of the 
primitive man that the sun was the cause of all life on the 
earth, and it was a natural conclusion that the ever-changing 
aspect of the sun, moon, planets and stars should be con- 
nected with the constant mutations in the lives of the in- 
dividuals of the race. Hence their gods and goddesses were 
identified with the heavenly bodies. The priests became 
astrologers. To be able to read the signs of the heavens 
seemed to be synonymous with understanding the happen- 
ings on earth, and consequently if this knowledge were to 
be utilized to predict what would take place in the future 
it was necessary to follow with the minutest care the vari- 
ous motions of the sun, moon, planets and stars and 
hence the priests became astronomers. These religious be- 
liefs caused a systematic study of the heavenly bodies; and 
as a result astronomy, the first and parent of all of the 
Sciences, had its birth. 


The movements of the sun and moon were symbols of law 
and order, and accordingly they were worshipped as gods. 
The motions of the planets, though more difficult to under- 
stand, seemed also the subject of order and not of caprice, 
and so likewise they were worshipped. Jupiter was the 
abode of the Babylonian god Mardok, and Venus was the 
home of the goddess of love, Ishtar. Saturn was identified 
with Ninib, Mercury with Nebo, and Mars with Nergal. 
In order to understand the manner in which the gods of the 
sun, moon and planets influenced the lives of men, it was 
necessary for the Babylonian priests to observe the position 
of these bodies, not only with respect to each other, but also 
with respect to the more prominent and easily recognizable 
fixed stars. In the case of the moon it was very necessary 
to note the time at which the new moon became visible as 
a crescent, its position in the heavens, and the angle made 
with the horizon by the line through the horns. (Even at 
the present supposedly enlightened age of the twentieth 
century A.D., we still hear of the " wet " moon, and the 
" dry " moon.) The changes of the moon were followed 
with meticulous care since each change afforded the op- 
portunity of interpreting some phase of human activity. 
And through forty centuries of the world's history some of 
these superstitions have survived in the popular belief that 
a change in the moon means a change in the weather, or 
that certain crops should be planted " in the dark of the 
moon." These observations of the heavenly bodies being 
thus gradually accumulated must have reached a mass of 
considerable proportion. For his interpretation of these ob- 
servations the astrologer rested on written records or on 
the recollection of what had happened under similar circum- 
stances in the past, but sometimes merely on the vagueness 
of association of ideas. 

The motion of the sun and moon among the stars near the 
ecliptic being carefully observed, it was but natural that 
the twelve full moons of the year should suggest the divi- 
sion of the circle of the year into twelve parts. We thus 
owe to the Babylonians the constitution and nomenclature 
of the twelve signs of the zodiac. Ages before AssurbanipaPs 


great library of tablets was constructed, the eighth month 
was known as " the month of the star of the Scorpion/' the 
tenth as " the star of the Goat/' while the twelfth was the 
month of " the star of the Fish of Ea." The convenience of 
the duodecimal system was thus early recognized. Each 
day was divided into twelve " double hours. " The ner of 
600, the sar of 3600 was formed from the soss, or unit of 
60. The circle was divided into 360, or six times sixty, 
with further divisions by 60 into minutes and seconds of 
arc, the hours being likewise divided into minutes and 
seconds of time. According to Miss A. M. Clerke, " In the 
Chaldean signs of the zodiac, fragments of several distinct 
strata of thought appear to be subdivided. From one point 
of view they shadow out the great epic of the destinies of the 
human race; again the universal solar myth claims a share 
in them; hoary traditions were brought into ex post facto 
connection with them; or they served to commemorate sim- 
ple meteorological and astronomical facts. " Astronomy 
was thus of very old standing in Babylonia. The principal 
astronomical work, called the Illumination of Bel was 
compiled for the library of Sargon of Akkad ; it was inscribed 
on 70 tablets, and apparently went through numerous edi- 
tions, one of the tablets being in the British Museum. It 
treats, among other things, on observations of comets, the 
pole star, the conjunction of sun and moon, and the motions 
of Venus and Mars. 

However, since the study of orderly motions of the 
heavenly bodies was primarily for the purpose of forecasting 
human events, it is not surprising to learn from modern 
researches that the early astronomical knowledge was very 
crude, with very little perfection in the observations. It 
is true that as early as the days of Khammurabi (2 100 B.C.) 
there were combinations of prominent groups of stars into 
outlines of animals and figures, yet there is no evidence that 
prior to 700 B.C. more than a small number of the constella- 
tions of the zodiac had found their place in the sky. 

The Babylonians had tables of squares and cubes calcu- 
lated from i to 60, they measured time by the sun-dial and 
by water clocks, were familiar with lever and pulley, and 


even possessed a lens turned on a lathe, but which was cer- 
tainly not used as a magnifying glass. Such proficiency in 
this early people would lead one to expect that though their 
observations may have been crude they had nevertheless 
perfected systems of moon calculation and planetary tables 
of a high order of excellence. The discovery of the Saros of 
223 lunations, so useful and important in the predicting of 
eclipses, has been attributed to the Chaldeans. They seem 
also to have been aware that Venus returns in almost ex- 
actly eight years to the same point in the sky and to have 
established similar relations of 46 years for Mercury, 59 
for Saturn, 79 for Mars and 83 for Jupiter. They were 
thus able to fix in advance the positions of the heavenly 
bodies, of such great importance in astrological lore. In 
fact it is generally thought that Hipparchus, the first great 
astronomer of the world, acquired much of his information 
from the Chaldeans, though unquestionably he verified 
all of their findings. The Babylonian sage Berossus founded 
a school about 640 B.C. in the island of Cos, and it is not 
impossible that Thales of Miletus was one of his pupils. 

In spite of the unbounded admiration which one must 
feel for the development of astronomy under the Chaldeans 
it seems much more plausible to believe that the greatest 
work of the Chaldeans was accomplished in the later years 
of their history, this taking place even after the fall of 
Babylon in 539 B.C., and in fact after the Greeks had in- 
vaded the valley of the Euphrates. This matter, and partic- 
ularly their acquaintance with the Saros, will be further 
discussed in dealing with the famous eclipse of Thales. 

The study of astrology thus securing a firm foundation 
under the Babylonians, it spread from them, directly or in- 
directly, to all quarters of the globe. It came to Greece 
about the middle of the fourth century B.C., and reached 
Rome before the opening of the Christian era. In India 
and China both astronomy and astrology were acquired from 
the Greeks, and are largely reflections of Greek theories 
and speculations. By the introduction of Greek culture 
into Egypt, astronomy and astrology were both cultivated 
in the land of the ancient Pharaohs during the Greek and 


Roman periods. The Arabs developed astrology and also 
astronomy, many of the names of the brighter stars being 
Arabic. In Europe as late as the i4th and isth centuries, 
astrologers had important positions at the royal courts 
and were consulted on all matters of great moment to the 
nation. With the revival of learning, and particularly after 
the development of the Copernican system which showed 
that the earth was but one of the planets, astrology became 
more and more pushed into the background, though as 
late as the iyth century horoscopes were cast by the great 
astronomer Kepler. Even in the United States of America 
and in England there are still many thousands of people 
who believe in the flat earth, and who have great confidence 
in the efficacy of horoscopes to correctly forecast the future. 
Traces of the Babylonian astrology are still found in our 
every-day language, for we still have such phrases as " I 
thank my stars," he was " born under a lucky star," or an 
" ill-starred undertaking." 

As already stated, the Babylonians had a calendar of 
twelve lunar months, and a seven-day week, the seventh 
day being a day of rest. The early Babylonians reckoned 
events from some great catastrophe, such as in our day, 
Chicago and its great fire, but later in their history they 
counted time by the years of the reigning king. There were 
several early dynasties, but the succession of rulers is fairly 
certain. In Assyria, on the contrary, a plan unique among 
the early peoples was followed in naming the year after 
officers, called eponyms, whose term extended for one year 
only. This " Eponym Canon," as it is called, began with 
the year 911 B.C. In the Almagest of Ptolemy there is a 
list of Babylonian, Assyrian and Persian kings who ruled 
in Babylon together with the years each of them reigned be- 
ginning with the accession of Nabonassar in 747 B.C., to 
the conquest of Babylon in 33 1 B.C. by Alexander the Great. 
This Ptolemaic Canon is confirmed by other Babylonian 
chronicles and also by the Assyrian Eponym Lists. 

These Assyrian tablets record three eclipses of the sun. 
The first, interpreted by Rawlinson in 1867, is referred to 
as follows: " In the Eponymy of Burgasole, Governor of 


Gozan, a revolt in Assur took place in the month Sivar, 
and the sun was eclipsed." To call attention to the im- 
portance of the event the ancient scribe drew a line across 
the tablet. This eclipse has been carefully investigated by 
a number of astronomers, by Ginzel, Airy, Hind and others, 
and the general conclusion is that Nineveh, where lived the 
Assyrian scribe, was just outside the path of totality, and 
that the greatest obscuration was about 9:47 in the morning 
of June 15, 763 B.C. The second eclipse mentioned in the 
Assyrian records took place in the reign of Esar-haddon, 
the son and successor of Sennacherib. This eclipse was 
partial in Assyria, but annular farther to the east, and the 
date was May 27, 669 B..C. The record of the third is a 
little uncertain, but it is probably the eclipse of June 27, 
66 1 B.C., during the reign of Assurbanipal. 

These are not the only eclipses observed by the Chaldeans, 
for in the Almagest of Ptolemy is a record of three eclipses 
of the moon observed in Babylon. The first of the three 
was a total eclipse on March 19, 721 B.C., while the other 
eclipses, partial only, were observed the following year, 
on March 8, and September i. As has been stated above, 
Ptolemy gave a list of the kings who reigned in Babylonia. 
These eclipses, of sun and moon, fix the dates of Eastern 
chronology with great exactness, the earliest date in the 
world's history to be thus accurately determined being the 
year 911 B.C. in Assyrian chronology. Earlier dates are 
known with increasing uncertainty. The various estimates 
of historians for the beginning of the first Egyptian dynasty 
differ as much as two thousand years! 

As a result of the Babylonian eclipses, it has been neces- 
sary to alter the chronology of the Bible by lowering the 
dates to the extent of twenty -four years. This will be re- 
ferred to later in connection with passages from the Holy 
Scriptures in Amos and Isaiah. 


In the neighboring country of Egypt excellent opportunity 
must have been afforded for a study of astronomy on ac- 


count of the clear skies of the East and from the intercourse 
which must certainly have taken place between the Egyp- 
tians and the Chaldeans of Asia Minor. Egyptian records 
accordingly have been carefully scrutinized to find what 
traces there are of early astronomical knowledge in the land 
of the Nile. And since a study of eclipses must concern 
itself with the beginnings of astronomy, a brief glance will 
be given here to early history in Egypt. 

The early Chaldean monuments are probably of an earlier 
date than those of Egypt, but the former are almost form- 
less piles of sundried brick, while the tombs and pyramids 
of the early Egyptian dynasties are many of them in excel- 
lent preservation. The history of Egypt is generally di- 
vided into five periods: 1 (i) The Ancient Empire, 3400- 
2160 B.C., comprising ten dynasties of kings; (2) the Middle 
Empire, with Thebes as capital, two dynasties; (3) the 
second Thebic, or New Empire (1588-1150 B.C.), compris- 
ing dynasties XVII-XX, separated from the Middle Em- 
pire by the Shepherd Kings of Arabia; (4) the Decadence 
period of six dynasties, 1150-324 B.C., which includes the 
Persian conquest in 525 B.C.; and (5) the Ptolemaic and 
Roman period, 324 B.C.-300 A.D. 

Little is known of exact dates before the Conquest by 
Alexander although repeated attempts have been made to 
determine them. The list of the succession of the Egyptian 
kings is known with certainty, but the chronology is uncer- 
tain for the following reasons: (i) The lengths of the in- 
dividual reigns are not known with precision, since the 
record seldom reached to the end of the reign, and did not 
allow for co-regencies. (2) Calculations on the probable 
length of a period leads to no trustworthy information. (3) 
Comparisons with other records, particularly with those of 
Babylonia and Assyria, give the most valuable knowledge e 
But the dates before 911 B.C. are not known accurately even 
in Assyrian chronology, and it is therefore not surprising 
that the date of the beginning of the XlXth dynasty should 
be estimated by competent authorities anywhere from 1490 
B.C. to 1315 B.C. (4) Of most interest to us here is the 

1 Hamlin, Encyclopedia Americana, 10, 12. 


astronomical information. The Egyptian number system 
was decimal, each power of 10 up to 100,000 being repre- 
sented by a different figure, on much the same principle as 
the Roman numerals. The day was divided into two periods, 
each of twelve hours, the beginning of the day being the time 
of sunrise. The year was divided into twelve months, each 
of thirty days, or a year of 360 days. As early as the Vth 
dynasty, the premature arrival of the seasons was noted, 
and as a consequence five complementary days were added 
making a year of 365 days. The extra days were counted 
either at the beginning or at the end of the year. The year 
was divided into three seasons, each month into three weeks, 
each of ten days. Since the season of agricultural growth 
depended more on the inundation by the Nile than on the 
motion of the sun, the first of the year was reckoned as the 
beginning of the rise of the waters of the river. As this took 
place with fair regularity, the Egyptians were thus furnished 
with a useful starting-point for their annual counting of 
time. It was noticed that the brilliant dog-star, Sirius, or 
Sothis, rose with the sun about this time (July 19), and as 
this star is so brilliant, refined astronomical measurements 
were not necessary in order to observe it. But the tropical 
year, the year of the seasons, is approximately a quarter of 
a day more than 365 days. (The tropical year is now known 
to be equal to 365.2422 mean solar days.) Hence the 
Egyptian calendar got astray one month every 121 years, 
or one year every 1461 years. This period, during which 
the New Year's day of the Egyptians traveled all round 
the calendar, was known to Greek and Roman writers 
in the first century B.C. Would that Julius Caesar when 
revising the Calendar the Julian being the basis of 
the Gregorian, or modern calendar had followed the sim- 
plicity of the Egyptian calendar, at least in keeping each 
month of a uniform number of days, and would that pride 
and jealously had not robbed poor February of two days! 
The modern calendar is very awkward and inconvenient, and 
there have been many attempts suggested for the purpose of 
revising it. Most of the modern revisions proposed call for 
an extra day in the year, or two extra days in leap years, 


which days are not to be included in the reckoning of any 
week or any month. Attention should be called to the fact 
that the five extra days of the Egyptian year recorded as 
early as the Vth dynasty were regarded as so peculiarly 
unlucky that there is not a single instance, among the many 
thousands of inscriptions brought to light, of any contract 
being entered into on any one of the five complementary 
days, nor has any event of importance ever been recorded 
as taking place on one of these unlucky days. 

The brilliant skies of Egypt favored the development of 
observations. It is generally believed that the Egyptians 
observed the motions of the heavenly bodies for the purpose 
of fixing the dates of their religious festivals. These times 
were probably noted by observing the times at which certain 
of the brighter stars arose at dawn just before the sun. For 
accomplishing this object no instruments were necessary. 

They also determined the hours of the night by observing 
the meridian passage of certain stars. What is thought to be 
the oldest type of observing instrument in the world, known 
to the Egyptians as merkhet, was probably utilized for this 
purpose. It consisted of a handle supporting a plumb bob 
and a reed which served as a sighting vane. By means of 
this and with the help of the ancient Egyptian horoscopus, 
or clock, the meridian could be laid out and time determined 
by observation. There have been preserved the titles of 
several temple books, which books apparently recorded the 
movements of the sun, moon and stars, but unfortunately, 
not a single one of these records themselves have survived. 
The Egyptians were of an intensely practical turn of mind, 
and they had little desire to acquire knowledge for the sake 
of the knowledge itself. If the information could be utilized 
for any practical use it was carefully cherished and pre- 
served, or if it could assist in their religious speculations 
it was then regarded of great value. If we are to judge of 
the accomplishments of the Egyptians by the written docu- 
ments alone, we must assume that they knew little and 
cared less for any branch of science. But then, there are the 
pyramids, and the great temples such as Karnak! Surely 
the architects and engineers who constructed such colossal 


monuments must have possessed knowledge of surveying 
vastly greater than is shown by any written record. The 
construction of the Great Pyramid apparently took place less 
than a century and a half after the elevation of the earliest 
piece of stone masonry known. Since these pyramids could 
have been erected only by means of great mechanical power, 
we are forced to the conclusion that the progress in the con- 
trol of such power was greater at the thirtieth century B.C. 
than at any other age of the world's history except the 
present. The pyramid was placed four square to the points 
of the compass, and leading to the sepulchral tomb there was 
a passage at an angle of 26 to the horizon, which has been 
thought furnished a corridor for observing the star which 
was then the pole star, a Draconis. With this assumption, 
an estimation has been made of the possible age of the pyra- 
mids. According to C. Piazzi Smyth, the passage in the 
pyramid facing the south was used for the observation of 
the meridian passage of the stars and planets, and it is even 
possible that the moon and sun were similarly observed. 
Proctor thinks that if the Egyptians had utilized an opaque 
screen with a small round hole in it at the southern end of 
this gallery they could even have observed sun-spots! 

It is plausible to believe that the Great Pyramid was pos- 
sibly designed to serve as an astronomical observatory, but 
the truth is that not a single competent Egyptologist sup- 
ports this view. As Breasted has pointed out, the careful 
orientation of the pyramid with its passage directed towards 
the north star is in keeping with the popular belief that the 
king was transmuted into a star, hence the shaft directed 
towards those stars in the sky which never set. 

According to E. W. Maunder, the symbol of the sun-god, 
so frequently found on Egyptian monuments, is sufficient 
evidence to prove that total eclipses of the sun were ob- 
served in Egypt. The general design of this symbol is a 
circle with striated wings to right and to left, and with 
a lesser extension in the downward direction, but with 
nothing above the circle. As pointed out by Maunder, 
the extensions to right and left call to mind and re- 
semble the form of the solar corona when the eclipse, oc- 


curring at sun-spot minimum, has the equatorial wings. 
This explanation is indeed plausible, and it is not impossible 
that it represents the truth. To use the coronal form for the 
symbol of the sun would have necessitated that the corona 
had been frequently observed by the Egyptians, and at more 
than one age. Such symbols could be adopted by a people 
only after slow awakening and continued observation. Con- 
sequently, if eclipses had been so frequently observed it 
would be rational to expect that records of some of them 
would have been preserved. Considering the frequency 
with which eclipses were observed in Babylon, it is surprising 
in the highest degree to find not a single reference any- 
where in Egyptian antiquity to an eclipse, either of the sun 
or of the moon. We are therefore forced to the conclusion 
that the Egyptians had little part in the development of the 
science of astronomy. 

The New Empire beginning with the XVIIIth dynasty 
marks the golden period of Egyptian life. This was the age 
of conquest with an empire stretching from the Euphrates 
to the fourth cataract of the Nile. Egypt had become a 
great military empire whose world power lasted from the 
early sixteenth to the twelfth century B.C., or over four hun- 
dred years. The capital was the once vast city of Thebes 
constructed on both sides of the river Nile, the first great 
monumental city built by man. On the east side of the river 
is the temple of Karnak which is nearly a quarter of a mile 
long and was nearly two thousand years in the course of 
construction. The obelisk of Queen Hatshepsut, the first 
great woman queen of history, is still standing. This is 
about one hundred feet high and weighs 350 tons. The 
first of the world's great generals was Thutmose III, who 
ruled for fifty years beginning about 1500 B.C., and on the 
walls at Karnak are inscribed the stories of his great ex- 

The recent discovery of the tomb of Tut-ankh-Amen has 
brought into sudden prominence that portion of Egyptian 
history immediately following the reign of Thutmose III. 
While little light so far has been thrown on the actual his- 
torical facts of the lives of Tut-ankh-Amen and his im- 












a * 


Pi TO 

PH ^ 

a* ^ 





mediate predecessors, the incomparable beauty of the 
objects of art found within the tomb bids fair to revolu- 
tionize the history of Egyptian art and literature. It is to 
the influence of the religious upheaval brought about by 
Khu-n-aten, the successor of Thutmose IV, and reputed 
father-in-law of Tut-ankh-Amen, that the revolution in 
artistic and literary expression is due. This monarch, 
known as the " heretic king/' physically a weakling, men- 
tally a poet and dreamer, revolted from the polytheistic 
religion of his forefathers and founded a new faith which 
deified the sun as the giver of all light and life. This 
worship, with Aten the sun as its visible symbol, had many 
points in common with Christianity and has been regarded 
as the most beautiful of the ancient religions. In his en- 
thusiasm for the new faith, however, Khu-n-aten allowed 
the material needs of his empire to go unheeded to such 
an extent that at his death the confusion and distress 
throughout Egypt was so great that his successor was forced 
to give way to the strong political power of the priests of 
the old religion, and Amen once more became the chief 
Egyptian god. It was in the shadowy period between Khu- 
n-aten and Rameses I that Tut-ankh-Amen's brief reign 

The oldest astronomical instrument in the world has re- 
cently been discovered by Breasted. The interest in this 
remarkable discovery is all the greater for the reason that 
it is believed that the instrument was actually made by no 
less a personage than King Tut-ankh-Amen himself. This 
object, a merkhet, consists of a rectangular strip of ebony 
wood 10 x i x Jr inches. Along each edge, extending from 
end to end, is an inscription which states that the instrument 
was made by King Tut-ankh-Amen, "with his two hands," 
as a restoration of a similar monument stolen from the tomb 
of his ancestor, Thutmose IV. Indeed, the tomb of the 
latter contains a remark in ink on the wall stating that his 
tomb had been restored (of course after robbery) by Harm- 
hab, who was practically the successor of Tut-ankh-Aijien. 

Apparently the ancient Egyptians were quite accustomed 


to despoiling the tombs of their ancestors; that of Thutmose 
IV had already been robbed during the reign of Tut-ankh- 
Amen while that of the latter king was soon to follow a 
similar fate. 

The XlXth dynasty began with Rameses I, whose reign, 
according to Breasted, commenced about 1315 B.C. This 
king reigned only two years, but he had planned and had 
started construction on the great colonnaded hall at Karnak, 
the greatest of such edifices ever constructed by man. It is 
338 feet long and 170 feet wide. There are one hundred and 
thirty-six columns in sixteen rows, those in the nave being 
higher than the rest. Rameses shared his throne with his 
son Seti I for one year, and Seti continued the gigantic 
plans of his father. In addition he built a splendid temple 
at Abydos, and a magnificent tomb in the Valley of the 
Tombs of the Kings. This galleried tomb is the very finest 
of its kind. The decorations are wonderfully charming, full 
of life and color though perhaps lacking some of the strength 
and character of earlier art. The outlines are forceful, the 
postures natural and the coloring beautiful. Egyptian art 
had reached the pinnacle. The Metropolitan Museum of 
Art of New York has been conducting operations at this 
magnificent tomb of Seti I. The photographs of the tomb 
examined with care give glimpses of the glories of the tomb 
and show some of the decorations which include the Lion 
surrounded by stars and the Bull. A careful scrutiny of 
photographs of this and other tombs appears to show that 
the stars connected with the Lion and the Bull in no way 
refer to the well-known zodiacal constellations of Leo and 
Taurus. In fact, the signs of the zodiac were not acquired 
from the Chaldeans until fairly late in Egyptian history, 
for there is no record of them until after the Ptolemaic 
period. If the Grand Pyramid had been in Babylon instead 
of Egypt it is safe to say that it would have been used as an 
astronomical observatory. The Egyptians were not inter- 
ested in systematized observations, and throughout the long 
years of their interesting history they contributed but little 
to the science of astronomy. 



IT IS safe to say that there is certainly one allusion in 
the Holy Scriptures to a solar eclipse, with one or two 
others possible. In Amos viii, 9, appear the words, 
" I will cause the sun to go down at noon, and I will darken 
the Earth in the clear day." The language is so unmistak- 
able and gives such a precise description of an eclipse of the 
sun that commentators have generally agreed that such a 
phenomenon must have taken place. This and other Scrip- 
tural references have been so fully and so carefully inves- 
tigated by Chambers in The Story of Eclipses, and by John- 
son, Eclipses, Past and Future, that a brief reference only 
will be given here. 

The date set down in the margin of certain Bibles oppo- 
site this passage in Amos is 787 B.C. Eclipses were visible 
in the neighborhood of Samaria in the years 791 B.C., 771 
B.C., 770 B.C. and 763 B.C. This last mentioned is the eclipse 
of Nineveh already described, an eclipse which was visible 
also in Samaria. There seems no doubt whatever that this 
is the eclipse predicted by Amos. It therefore becomes 
necessary to lower by twenty-four years the date given in 
Amos; and if this is done there is a ready explanation of 
the story in the life of King Hezekiah, as given in II Kings 
xx, ii. The Old Testament thus reads: " And Isaiah the 
prophet cried unto the Lord: and he brought the shadow 
ten degrees backward, by which it had gone down in the 
dial of Ahaz." Ahaz was the father of Hezekiah, and the 
" dial " was probably a sun-dial similar in construction to 
the ancient sun-dials of masonry still existing at Benares 
and Delhi in India. 

The Biblical peoples at this period of their history were 
divided into two. In the North, the land was rich and fer* 



tile and great prosperity and wealth abounded. The people 
Hved in large towns, they dressed and lived extravagantly. 
Moreover these people of Israel had adopted the gods of the 
Canaanites, each town in fact having its own god, or " baal." 
To the South, the land was poor, encroaching as it did on the 
desert, and the people of Judah had to struggle for their 
existence. The only large town was Jerusalem. The Jews 
were still faithful to the Hebrew God, Jehovah. The shep- 
herd Amos being a devout follower of Jehovah made a 
journey to Israel, and thundered his denunciations against 
their many gods, and against their gaudy clothes, their licen- 
tiousness and harsh treatment of the poor. 

About this time the Hebrews were beginning to learn to 
write, and they were abandoning the clay tablet and the 
cuneiform inscription of the Assyrians and Babylonians. 
They wrote on papyrus and with pen and ink, which method 
of writing they had acquired from their former masters, the 
Egyptians. However, they borrowed the first alphabet the 
world had developed, that of Phoenicia. The papyrus rolls 
written by Amos and other historians have descended to us 
as the Hebrew Scriptures. On account of the evil lives of 
the people of Israel, Amos predicted their destruction. 
Damascus was taken by the powerful Assyrians in 732 B.C., 
and thus being unprotected on the north, Samaria, the capi- 
tal of the kingdom of Israel, was captured in 722 B.C., the 
Israelites being taken away captive and their nation de- 

Shortly before 700 B.C. arose the great prophet Isaiah. 
In one great oration after another he exhorted the people of 
Jerusalem to believe in their God, Jehovah, and not be dis- 
mayed for He would deliver them from the hosts of the 
powerful Assyrian, Sennacherib, who was then threatening 
to batter down the city. The king of Jerusalem, Hezekiah, 
was sick unto death, and the same tragic fate as had befallen 
Damascus and Samaria seemed to be awaiting the people of 
Jerusalem. In answer to a prayer of Hezekiah for recovery 
the prophet was sent to him with this message: " Thus 
saith the Lord, the God of David thy Father, I have 
heard thy prayer, I have seen thy tears: behold I will add 


unto thy days fifteen years . . . and I will defend this city, 
and this shall be a sign unto thee from the Lord, that the 
Lord will do this thing that He hath spoken. Behold, I will 
bring again the shadow of the Sun, which is gone down in 
the sun-dial of Ahaz ten degrees backward, by which degrees 
it had gone down." (Isaiah xxxviii, 5-8.) 

There was a large partial eclipse of the sun visible in 
Jerusalem on January n, 689 B.C. If we accept the correc- 
tion of twenty-four years derived from the record of the 
eclipse of Nineveh, then according to Chambers (loc. cit.\ 
this eclipse of Jerusalem will completely satisfy the Biblical 
narrative at all points. History tells that a pestilence from 
the marshes of the Nile caused great havoc in the army of 
Sennacherib, and thus was Jerusalem miraculously spared. 
See also in this connection, II Kings, xix, 32-37. About a 
century later the Hebrews rejoiced over the destruction of 
Nineveh in 606 B.C. But unfortunately, the days of Jeru- 
salem were numbered. The Chaldeans followed Assyria as 
masters of Palestine, Jerusalem was destroyed in 586 B.C. 
under Nebuchadnezzar, and the people were carried away 
into exile in Babylon. 

The dates herewith given in Babylonian and Assyrian 
history admit of no uncertainty since they are determined 
by eclipses. As a consequence, the dates appearing in the 
Bible must be altered to fit those of verified history. It 
would be interesting to follow the history of the Hebrews 
through the lives of Jeremiah and other prophets but 
space forbids. The reader is referred to Breasted, Ancient 
Times, a History of the Early World, and also to many 
other books on this period of history which is of the greatest 
of interest to all Christian peoples. 


It will be possible to give an account here of but a few of 
the eclipses referred to in the Classics, and we shall begin 
with Homer. On the day of the slaughter of the suitors, 
there is a passage in the Odyssey (v. 351-357), which prob- 
ably refers to an eclipse of the sun. The lines run: 


" Ah, wretched men, what evil is this that you suffer? 
Shrouded in night are your heads and your faces and your 
knees beneath you; kindled is the sound of wailing, bathed 
in tears are your cheeks, and sprinkled with blood are the 
walls and the fair rafters. And full of ghosts is the porch 
and full the court, of ghosts that hasten down to Erebus 
beneath the darkness, and the Sun has perished out of 
heaven, and an evil mist hovers over all." 

According to Fotheringham (loc. aY.), the words, " The 
Sun has perished out of heaven/' must refer to a total 
eclipse, and in fact most commentators, including Plutarch 
and Eustathius, agree in this interpretation. Apparently, 
the eclipse was contained in the legend as it descended to 
Homer, though naturally it is impossible to fix the exact 
date of the eclipse. 

The next eclipse is one referred to by the Greek poet 
Archilochus, a portion only of whose work has come down 
to us. The lines are: " Zeus, the father of the Olympic 
Gods, turned mid-day into night, hiding the light of the 
dazzling Sun; and sore fear came upon men." According 
to the investigations of Oppolzer and Cowell, the poet 
must have witnessed the eclipse of April 6, 648 B.C., the 
eclipse being total about 10 A.M. at Thasos and in the north- 
ern part of the Aegean Sea. We know that the poet spent 
part of his time at Thasos, and if his statement actually 
gives the description of an eclipse, it thus furnishes the 
earliest date in Grecian chronology to be fixed with cer- 
tainty. The early accepted dates must accordingly be re- 
duced by fifty years, but no surprise need be felt as they 
were known only with great uncertainty. 


The most celebrated eclipse of all history is that con- 
nected with the name of Thales of Miletus who lived from 
640 to 546 B.C. He was the founder of Greek astronomy, 
geometry and philosophy. He was regarded by the Greeks 
with great veneration, and he was the chief of the seven 
" wise men," a reputation which rested not only on his 


scientific eminence but also on his political sagacity. Ac- 
cording to the historian Herodotus (i, 74), the reference to 
the eclipse is as follows: " There was war between the 
Lydians and the Medes for five years, each won many vic- 
tories from the other, and once they fought a battle by 
night. They were still warring with equal success, when it 
chanced, at an encounter which happened in the sixth year, 
that during the battle the day was turned into night. Thales 
of Miletus had foretold this loss of daylight to the lonians, 
fixing it within the year in which the change did indeed hap- 
pen. So when the Lydians and Medes saw the day turned 
to night they ceased from fighting, and both were the more 
zealous to make peace." The truce concluded was cemented 
by a double marriage, " for/' adds the historian, " without 
some such strong bond, there is little of security to be found 
in men's covenants." The same eclipse is referred to by 
both Pliny and Cicero. 

The narrative by Herodotus contains two statements, 
that the eclipse was total (a " night battle "), and that the 
eclipse was predicted by Thales. It is in regard to the 
second fact that the controversy has arisen among* astrono- 
mers. Various dates for the eclipse have been assigned, 
from 625 B.C. to 583 B.C. But after careful researches of 
many competent authorities Airy, Hind, Zech, Hansen, 
Ginzel, Newcomb, Cowell, Fotheringham and others the 
date of May 28, 585 B.C., has been fixed, the eclipse taking 
place in the afternoon. The general consensus of opinion is 
that Thales was probably familiar with the Chaldean Saros 
of 223 lunar months or 18 years n days. In fact, by some it 
has been thought (see, George Smith, Assyrian Discoveries} 
that the Babylonians were actually making use of the Saros 
for the prediction of eclipses. Herodotus does not claim that 
the day and hour of the eclipse were predicted nor yet the 
locality where it was to be visible, these being much more 
difficult problems and incapable of being solved by Thales. 
The eclipse was predicted only " within the year." It is 
therefore not necessary to credit Thales with extraordinarily 
abnormal powers, or to assume that he was in possession of 
information which was known first to Hipparchus more than 


four hundred years later. Apparently the Greek " wise 
man " utilized the eclipse of May 17, 603 B.C., for the pur- 
pose of his prediction. 

Probably the greatest difficulty of all concerning the iden- 
tification of this eclipse of Thales has been to know just how 
much credence could be placed in the account of Herodotus. 
Some of the ancient writers go so far as to accuse him of 
" intentional untruthfulness." Modern critics while not go- 
ing to this extreme, none the less feel that his love of effect 
and his loose and inaccurate habits of thought make of him 
an attractive writer, but a poor historian. A case in point 
might be cited, the eclipse which is described by Herodotus 
(vii, 37) as follows: " At the first approach of spring, the 
army quitted Sardis and marched towards Abydos; at the 
moment of its departure the Sun quitted its place in the 
heavens and disappeared, though there were no clouds in 
sight and the day was quite clear; day was thus turned into 
night. " We are told that " As the King was going against 
Greece, and had come into the region of the Hellespont, 
there happened an eclipse of the Sun in the East; this sign 
portended to him his defeat for the sun was eclipsed in the 
region of his rising, and Xerxes was also marching from that 
quarter." The generally accepted date of the battles of 
Thermopylae and Salamis is 480 B.C. No eclipse occurred in 
the spring of that year. Many attempts have been made to 
find an eclipse that would harmonize with the statement of 
Herodotus, but all of no avail, unless we are willing to 
make changes either in the date or in the narrative. It 
seems necessary to conclude that the eclipse incident adds 
greatly to the attractiveness of the story, but that it is 
not history. Unfortunately, even at the present day, 
authors still try to embellish their stories with celestial 
phenomena, but do not stick too closely to the facts. One 
of the " best sellers " appearing recently in the field of fic- 
tion had in it no less than four astronomical blunders. 

The next eclipse to be mentioned was seen in Athens in 
B.C. 431, August 3. It is described by Thucydides (ii ? 28) 
who states that during the Peloponnesian War " things for- 
merly repeated on hearsay, but very rarely confirmed by 


facts, became not incredible, both about earthquakes and 
eclipses of the Sun which came to pass more frequently than 
had been remembered in former times." An eclipse occurred 
in the first year of the war, " in the same summer, at the 
beginning of a new lunar month (at which time alone the 
phenomenon seems possible), the Sun was eclipsed after 
midday, and became full again after it had assumed a cres- 
cent form and after some of the stars had shone out" The 
account, differing from the flowery but uncertain language 
of Herodotus, is clear and definite, and refers to an eclipse 
which evidently was not total at Athens where Thucydides 
was supposed to be. The only difficulty in fixing the eclipse 
has arisen from the last part of the statement, that stars 
were visible. Venus was 10 distant and undoubtedly must 
have been readily seen, but what other star, or stars? The 
eclipse in B.C. 431 was seven-eighths total. To make a long 
story short, it sufficeth to say that at the eclipse of April 8, 
1921, which was eight-ninths total at Oxford, the star Vega 
was readily seen. It must be concluded, therefore, that the 
account of Thucydides is strictly in accord with the facts 
and that the stars Venus, Vega and probably Jupiter 43 
distant were seen. 

This same eclipse has an interesting anecdote connected 
with it which has been narrated by Plutarch in his Life of 
Pericles who was the commander of the Grecian naval 
forces. " The whole fleet was in readiness, and Pericles on 
board his own galley when there happened an eclipse of the 
Sun. The sudden darkness was looked upon as an unfavor- 
able omen, and threw the sailors into the greatest consterna- 
tion. Pericles, observing that the pilot was much astonished 
and perplexed, took his cloak, and having covered his eyes 
with it, asked him if he found anything terrible in that, or 
considered it a bad presage? Upon his answering in the 
negative, he asked ' Where is the difference, then, between 
this and the other, except something bigger than my cloak 
causes the difference? ' " 

Another eclipse recorded by Thucydides (iv, 52) has been 
identified with the annular eclipse of March 21, 424 B.C. 

The last of the eclipses described by Thucydides (vii 50) 


is one of the moon, and it has a tragic story connected with 
it. The Athenian fleet and army were engaged in an attack 
on Syracuse, the fleet being under the command of Nicias, 
and Demosthenes had arrived with large reinforcements 
for the army. The latter had failed in his purpose of cap- 
turing a wall which the Syracusians had thrown across the 
Athenian lines, and as a result of this failure it had been 
decided to withdraw the whole Athenian forces. On the 
night of August 27, 413 B.C., the whole force had embarked 
and was ready to flee. Before a start was made, however, 
an eclipse of the moon took place which became total. 
Nicias was greatly terrified, for it seemed easy to under- 
stand the cause of an eclipse of the sun, but very difficult 
to make out " how the moon, when at the full, should sud- 
denly lose her light and assume such a variety of colors." 
The soldiers and sailors added their terror to that of the 
commander and as a result the soothsayers were consulted. 
The advice was to remain for thrice nine days, but alas, 
before that time the Athenians had been utterly routed, and 
Nicias and Demosthenes were both executed. 

Plutarch records an eclipse of the sun which took place 
July 13, 364 B.C., and both Plutarch and Pliny give an ac- 
count of a total eclipse of the moon September 20, 331 B.C., 
eleven days before the victory of Alexander over Darius at 
Arbela in Assyria. 

Another interesting eclipse of the sun is connected with 
the name of Agathocles, the tyrant of Syracuse. The narra- 
tives of the two historians Justin and Diodorus Siculus agree 
so well that the astronomer has been able to fix exactly the 
circumstance of the eclipse. In 310 B.C. Syracuse was again 
besieged, this time not by Athens, but by Carthage. On 
the evening of August 14, Agathocles slipped out of the har- 
bor with his whole fleet of sixty ships. He was pursued by 
the Carthaginians, who abandoned the chase at night-fall. 
According to the historical accounts, there was an eclipse of 
the sun on the following morning, and as a consequence day 
was turned into night, and stars appeared everywhere in the 
firmament. The soldiers were at first afraid, believing the 
eclipse to be an ill omen, but their fears were quickly calmed 


by their commander who, according to Justin, told them that 
eclipses " always signify a change of affairs, and therefore 
some change was certainly signified, either to Carthage, 
which was in such a flourishing condition, or to them, whose 
affairs were in a very ruinous state." Apparently, these 
words of Agathocles seemed to have inspired his men that 
an eclipse might even be an omen for good. At any rate, 
he was able to make good his escape from the fleet of the 
Carthaginians, and after six days and nights he succeeded 
in landing on the coast of Africa and in devastating the 
territories of proud Carthage. 

The readiness with which Agathocles and Pericles quieted 
the fears of the sailors calls to mind the genius of Columbus 
who ? on more than one occasion, had to quell mutinies among 
the members of his little fleet, they being greatly perturbed 
on discovering that the compass needle did not point to the 
true north as given by the pole star. 

Mention of only one other eclipse before the beginning 
of the Christian era will be made here. An eclipse is gen- 
erally regarded to have taken place when Julius Caesar was 
crossing the Rubicon on his triumphal return to Rome from 
Gaul. But it seems to be necessary to class this eclipse 
along with that connected with the name Xerxes, for it 
failed to occur. 

In the investigations into the authenticity and accuracy of 
the early eclipses, the greatest difficulty has arisen from the 
vagueness and uncertainties of the account of the chronicler. 
References to eclipses are couched in such hazy language 
that the account might have been equally well applied to 
some other rare phenomenon. Writers of literature in all 
ages seem to prefer flowery to straightforward language and 
employ imagery rather than fact, refusing to " call a spade, 
a spade." For the astronomer it is impossible to take one 
interesting eclipse and decide on the pros and cons in favor 
of this or of that interpretation. Eclipses depend on the 
motions of the sun and moon, and the history of all eclipses 
must make a harmonious story and not a disjointed one. 
And so the eclipse of Agathocles must be considered in 
reference to that of Thales, and that in turn to the eclipse 


of Nineveh in 763 B.C. The astronomer is not a dealer in 
necromancy nor does he acquire information about the mo- 
tions of the sun and moon by mystical incantations, or by 
divination. The work of the scientist is as far removed as 
possible from the mountebank and charlatan. Scientific 
knowledge is acquired only by the patient study of all that 
has been accomplished in the past, and happy and fortunate 
is he who can make a correct interpretation of the past! 
The manner in which the story of eclipses has added to 
astronomical knowledge will be treated in a future chapter. 
Those who wish additional information regarding the early 
eclipses would do well to consult Encyclopaedia Britannica, 
article on Eclipses by Simon Newcomb, but specially the 
monumental work by Oppolzer, Canon der Finsternisse, 
which gives a record of no less than 13,200 eclipses, of which 
5,200 are of the moon and 8,000 of the sun. These eclipses 
are the ones which have taken place since 1208 B.C., over 
three thousand years ago, or which will take place before 
the end of the year 2162 A.D., two hundred years in the fu- 
ture. In addition to the tables of the astronomical elements 
of each eclipse, 160 charts are furnished giving the location 
on the earth's surface where each total and annular eclipse 
of the sun will be visible. It should, however, be pointed 
out that the eclipse tracks could not be located by Oppolzer 
on the maps with the greatest of refinement, since it was 
manifestly impossible to calculate the location of each path 
for more than three positions, sunrise, noon and sunset. 
Consequently, it need not be a matter of surprise, par- 
ticularly with the early eclipses, that the tracks may at times 
be incorrectly placed by as much as a hundred miles. 



AFTER the destruction of Babylonia and Assyria, 
astronomy migrated from its Chaldean home to 
Greece. As early as the fifth century B.C., attempts 
were made to demonstrate some of the practical uses of 
astronomy in the regulation of time and particularly in the 
arranging of the calendar. The Greeks had inherited from 
the Chaldeans a calendar founded on the lunar month, but 
since the month is 29.5 days in length, attempts to bring 
the seasons and the calendar into harmony ended always in 
hopeless confusion. The astronomer Meton (born about 
460 B.C.) made the discovery that 19 solar years were very 
nearly equivalent in length to 235 lunar months the 
difference, in fact, being only about half an hour. The 
Metonic cycle, as it has since been called, has been of the 
very greatest service since its discovery so many centuries 
ago. It should not be confounded with the interval called 
the Saros, since the Metonic cycle is not used for the predic- 
tion of eclipses. After nineteen years, new and full moon 
and the various phases are repeated with a considerable 
degree of exactness. This cycle is still used for finding the 
day on which Easter will fall. 

The Chaldeans had confined their energies to the making 
of astronomical observations, but their successors, the early 
Greek philosophers, made practically no observations whose 
record is worth keeping. On the other hand they were much 
interested in inquiring into the causes of things, and ac- 
cordingly it may be said that the science of astronomy had 
its birth with the Greeks. We are already familiar with 
Thales and part of his work. To Pythagoras we owe the 
doctrine of the " music of the spheres/' a notion which has 
descended even to the present day. To Plato (428-347 



B.C.) astronomy owes but little, he even going so far as 
to decry, as degrading, the observation of the heavenly 
bodies. Far different, however, was the case with Aristotle 
(384-322 B.C.) who appears to have collected and systema- 
tized the best thought of the time regarding astronomy. 
He believed that all of the heavenly bodies, including sun, 
earth and moon, were spherical; he taught that the moon 
had no light of its own but shone by reflected sunlight and 
he explained its phases. Most important of these conclu- 
sions was the proof of the spherical shape of the earth. 
This, as Aristotle demonstrated, depends on two facts: first, 
that the earth casts a shadow and that an eclipse of the 
moon is due to the passing of the moon into this shadow; 
and second, that the shadow being circular in outline was 
proof that the body casting the shadow, the earth, was cir- 
cular, at least in the section turned towards the moon. 

After the time of Aristotle, Greek astronomy moved its 
home to Alexandria, and there under the protection of suc- 
cessive Ptolemies the great museum was erected and a 
library and an observatory were incorporated as important 
adjuncts. It is not the purpose here to trace the develop- 
ment of astronomy except as it progressed through the study 
of eclipses. The greatest astronomer of antiquity, in fact 
the most famous up to the time of Newton, was Hipparchus. 
He was not of Alexandria, though he probably visited the 
city and may have made some of his observations there. 
Little is known of his birth. This may have taken place in 
Bithynia or in Rhodes, though it is known that he had an 
observatory in the latter place and there did most of his 
work. We know of him mainly through the astronomer 
Ptolemy, since all of the original works of Hipparchus have 
been lost with the exception of one important book. His 
great fame rests upon three pedestals: (i) The invention 
of trigonometry, (2) the making of an extensive series of 
careful observations, (3) the comparison of his own with 
earlier observations. As a result of his exact and methodical 
work, he discovered the precession of the equinoxes, he 
made the first catalogue of the positions of stars (to the 
number of 1080), and he contributed greatly to the theory 


of the motion of the sun and rnoon, and, as a direct con- 
sequence, to the subject of eclipses. 

Hipparchus was the first to notice that the seasons were 
of unequal length, the time from vernal equinox to summer 
solstice being 94 days, while the summer season was only 
92 1 days in length. Spring and summer together are seven 
days longer than fall and winter. Hipparchus gave the 
correct explanation of this fact, viz: that the earth is not 
always at a constant distance from the sun, being removed 
from the center of the orbit (which was then regarded as 
circular). At the present day this is readily verified by the 
change in the angular diameter of the sun, but this con- 
firmation in Hipparchus' time would have required measure- 
ments too exact for his crude instruments. By noting the 
time it takes the moon to pass through the earth's shadow 
at the time of a total eclipse of the moon, Hipparchus was 
able, by a method due to Aristarchus, to obtain a value 
of the relative sizes of earth and moon, and also to esti- 
mate the distance of the moon to be 59 times the terrestrial 
radius, which is not very far from the truth. 

The motion of the moon was much more difficult to in- 
vestigate than that of the sun. Little observation is neces- 
sary to note that the moon moves eastward in the sky, 
changing her position among the stars by her own diameter 
in approximately an hour. A revolution of the moon from 
a star to the same star again is known as a sidereal month, 
and its length is 27.3 days. In this interval the earth has 
proceeded in its orbit about the sun, and consequently the 
moon, as seen from the sun, requires more than two ad- 
ditional days to swing into line with the earth. The in- 
terval of time from new moon to new moon, or from full 
moon to full moon, is the ordinary month of 29.5 days, called 
the synodic month. Further observation shows that the 
sun is always in the ecliptic, in fact, the sun's apparent mo- 
tion determines the ecliptic, whereas the motion of the 
moon, although approximately in the ecliptic, is neverthe- 
less inclined at a small angle, which Hipparchus for the 
first time fixed as at an angle of 5. The moon's path thus 
crosses the ecliptic at two points called the nodes. If the 


moon's path were exactly in the plane of the ecliptic there 
would be two eclipses each month, once at the time of new 
moon when the moon would come between the earth and 
the sun, bringing about an eclipse of the sun; and again at 
full moon, when the moon would pass into the shadow cast 
by the earth and be itself eclipsed. But as the moon's orbit 
is inclined to the ecliptic, it is manifest that eclipses, either 
of sun or moon, can occur only when the moon is near the 
plane of the ecliptic, or, in other words, near its node. If 
the length of time it takes the moon to return to its node 
were the same as the sidereal month, it is evident that the 
moon's node would be fixed in space and would have no 
motion relative to the stars. The length of the month deter- 
mined by the return of the moon to the node is called the 
nodical, or the draconic month. The meaning of the first 
designation is evident. But why the second? It is not diffi- 
cult to trace it back to the ancient belief that the sun was 
swallowed by a dragon at the time of an eclipse, and this 
superstition is therefore the reason why the month on 
which eclipses depend should be called draconic. Indeed 
the symbols that are universally used by astronomers to de- 
note the ascending node ( & ) and the descending node ( W ) 
are generally supposed * to represent the head and tail of the 
dragon. Hipparchus found that the moon's nodes were not 
fixed, but that they completed a revolution in the plane of 
the ecliptic from east to west in about 19 years. A fourth 
kind of month was likewise known to Hipparchus, the length 
of a revolution from the position of perigee to perigee in the 
moon's orbit. By making use of the eclipses observed by 
the Chaldeans, Hipparchus was enabled to obtain greatly 
improved values of the lengths of the various months. The 
values that follow are those furnished by the recent investi- 
gations of E. W. Brown: 

Synodical = 

29.0530588 - 

2Q d 

I2 h 

44 rn 


S 8 

Sidereal = 


321661 = 






Anomalistic = 


554550 = 






Nodical = 


2I222O = 






See Berry, A Short History of Astronomy, p. 48. 


The lengths of the different kinds of year will be inserted 
here for future reference, the values being Newcomb's: 

Tropical (ordinary) = 365^242199 
Eclipse = 346. 620031 

Since the motion of the moon's node is in the direction 
to meet the oncoming sun, the interval of time from node 
to node (the eclipse year) is less than the interval from 
vernal equinox to vernal equinox, the ordinary year on 
which the seasons depend. From the above figures, the 
eclipse year is 18.62 days shorter than the tropical year. 
Since the sun and earth are always in the plane of the 
ecliptic, all that is required to permit the prediction of an 
eclipse of the sun is to find at the time of the 
new moon (when alone a solar eclipse can take place) 
whether the moon is sufficiently near enough to the plane 
of the ecliptic for her to come between earth and sun. And 
as an eclipse of the moon can happen only at full moon, 
similar investigations will furnish the means of predicting 
lunar eclipses. With the modern values given above, these 
predictions can be carried out remarkably easily. Let us 

19 eclipse years = 6585^7806 

223 ordinary months = 6585. 3211 
247 nodical months = 6585. 3572 
239 anomalistic months = 6585. 5374 

The integral parts of these four quantities are the same, 
the second being the Saros (meaning " repetition ") which 
amounts to 18 years n days, if only four leap years inter- 
vene, or 1 8 years 10 days if the 2gth of February has come 
five times. The eclipse of June 8, 1918 was a repetition 
of that of May 28, 1900; the eclipse of September 10, 1923 
followed that of August 30, 1905 by the interval of one 
Saros. The author observed the 1900 eclipse in Georgia 
and that of 1918 in Oregon. It was necessary for him to 
travel to Spain to witness the 1905 eclipse and to California 
for the one of 1923. Succeeding eclipses in the Saros are 
visible each time from a location on the earth's surface 
farther west. The reason for this, and the causes of 
further differences in eclipses which follow each other are 



readily explained from the fact that the numbers given 
above are not exactly equal to each other but differ in the 
decimal parts. Before going into this in detail, it will be 
well to take up some of the geometrical features useful in 
the prediction of eclipses. 

FIG. i 

The length of the shadow cast by earth or moon can be 
readily found. S is the center of the sun, E that of the 
earth, and M and m the moon's center at new and full 
moon, respectively. A shadow cone envelopes the sun and 
earth, its vertex being at T. If the moon in her orbit passes 
wholly into the shadow there is a total eclipse oj the moon. 
If the line ED is drawn parallel to TEA, then it is seen that 
the triangle DSE and BET are similar, and hence 

E T 
E B 

S I) 

But SD = SA-AD=R-r, if R and r denote the radii of 
sun and earth respectively. SE is the distance of sun to 
earth which equals A. Hence the length of the shadow, 


E T = 


But R, the radius of the sun, is equal to 109.5 times the 
radius of the earth, r, and accordingly, the length of the 
earth's shadow is equal to 



Assuming the average value of the distance from sun to 
earth as 92,900,000 miles, then the length of the earth's 
shadow is on the average 857,000 miles, the distance of the 
earth to moon being much less than this, or 238,840 miles. 
In quite a similar manner it is possible to find the length 
of the shadow cast by the moon, by taking the moon instead 
of the earth, and assuming C as the length of the moon's 
radius, and D the distance from centers of sun to moon. 
The length of the shadow cast by the moon is equal 



But R and C ? the radii of the sun and moon, are each con- 
stant, and substituting the values it is found that the length 
of the moon's shadow is 3-9^75-5 of the moon's distance from 
the sun, or approximately one four-hundredth part of the 
distance. Under the conditions given a total eclipse 
of the sun, the distance from sun to moon, D, is 
equivalent to the difference between the distances from 
sun to earth and from earth to moon. On account of 
the fact that the orbits of earth and moon are both 
ellipses and not circles, the distance D will vary consider- 
ably and the length of the moon's shadow will change pro- 
portionally. Under average conditions, the length of the 
shadow measured from the center of the moon is 231,650 
miles, but this may vary about 4000 miles each way from 
the mean, or from 235,700 miles as a maximum to 228,120 
miles as a minimum. Since the mean distance from center 
of earth to center of moon is 238,840 miles, or from moon's 
center to earth's surface 234,900 miles, it is seen that under 
average conditions the moon's shadow is not long enough 
to reach the earth's surface. The distance D will have its 
greatest value when the earth is at its greatest distance from 
the sun, at aphelion, which takes place about July i of each 
year, and the moon at its least distance from the earth, 
or at perigee. Owing to the revolution of the moon's line of 
apsides, perigee may come during any month of the year. 
The distance D is least when the earth is at perihelion 
(January i) and the moon at apogee. On account of the 


elliptical character of the moon's orbit, the distance from 
center of earth to center of moon varies, from approximately 
253,000 miles as a maximum to 221,600 miles as a min- 
imum. The latter distance is 217,650 miles from the earth's 
surface. Since the shadow cast by the moon may be 235,700 
miles in length, its vertex therefore at times may extend 
18,000 miles beyond the earth's surface. Under these con- 
ditions the area of the moon's shadow cone intercepted by 
the earth will be a maximum, and when all conditions are 
most favorable, the diameter will be 163 miles. Such condi- 
tions are possible when the total eclipse is visible about July 
i, with the moon at perigee, the eclipse being visible at the 
earth's equator at noon. The diameter of the moon's 
shadow path intercepted by the earth may vary from its 
maximum value to a vanishing width when the shadow is 
just long enough to reach the earth's surface. The average 
width is approximately one hundred miles. Inside this 
shadow cone an observer will see the light of the sun totally 
cut off; outside of it the sun will be partially eclipsed. 

Since also the distance may be as great as 253,000 miles 
from the earth's center, or about 249,000 miles from its 
surface, while the moon's shadow may be only 228,120 miles 
in length, the shadow may fall short of the earth's surface by 
more than 20,000 miles. Under these conditions an ob- 
server on the earth located on the axis of the shadow pro- 
duced, would see an annular eclipse and not a total one, a ring 
or annulus of light appearing round the edge of the moon 
when the eclipse is central. The diameter of the so-called 
negative shadow may be as great as 230 miles. Since the 
maximum width of the shadow causing a total solar eclipse 
is only 163 miles, it is evident that the number of annular 
eclipses are more frequent than total only two out of 
five central eclipses being total. 

The work of the observational astronomer has been 
greatly hampered by the moon which illuminates the sky 
and renders it impossible to see or to photograph the faint- 
est stars in the telescope. As already has been noted, the 
moon has furnished much labor and vexation of spirit to 
the mathematical astronomer. Yet if our satellite were 


banished from the sky, what a dreary old world it would 
be to poets and lovers without their " inconstant moon "! 
If the diameter of the moon were decreased a scant hundred 
and forty miles (less than seven percent), a total eclipse 
of the sun would be impossible! How thankful solar as- 
tronomers should be for the strange coincidence that the 
angular diameter of the moon is about equivalent to that 
of the sun and that the diameter of the moon is not ten 
percent less than it is ! The greatest possible excess of the 
angular radius of the moon over that of the sun is only 
i' 19". 

Knowing the diameter of the shadow intercepted by the 
earth it is easy to find the duration of totality. For this 
purpose it is necessary to find the speed of the shadow 
over the earth. The moon advances in her orbit approxi- 
mately her own diameter in an hour, or more exactly, about 
2100 miles per hour. On account of the great distance of 
the sun, this is the speed that the moon's shadow passes 
across the earth regarded as a whole, but on account of the 
rotation of the earth on its axis the velocity of the moon's 
shadow over the earth's surface is far different from 2100 
miles per hour. The diurnal rotation causes an observer 
at the equator to complete a circuit of nearly 25,000 miles 
in twenty-four hours which is at the rate of 1040 miles per 
hour. Away from the equator, farther north and south in 
latitude, the speed is progressively less at greater and 
greater distances from the equator, being in fact 1040 miles 
per hour multiplied by the cosine of the latitude. At 30 
north and south an observer is carried 900 miles per hour, 
at 45 only 735 miles, while at 60 this is reduced to 500 
miles. The moon moves in her orbit from west to east and 
her shadow in consequence traverses the earth in this direc- 
tion also. This being likewise the direction that an observer 
is carried by the earth's rotation, the moon's shadow path 
travels with respect to the earth's surface the difference of 
the two speeds, which is 1060 miles at the equator, and 
1600 miles at latitude 60. 

The above values refer only to the condition that the 
eclipse occurs on the meridian for the observer in question, 


or in other words, at noon. When the eclipse takes place 
near sunrise or sunset, the observer being turned not di- 
rectly towards the sun and moon, the projection of the 
moon's orbital velocity along the earth's surface may be 
very large. The slowest speed that the moon's shadow 
can have over the earth's surface is 1060 miles per hour, 
when the eclipse occurs at noon and at the equator, and the 
speed under conditions of higher latitudes and eclipse at 
sunrise or sunset may be as great as 4000 or even 5000 
miles per hour. With the conditions causing maximum 
width of shadow of 163 miles, an eclipse of the sun may 
have the maximum possible duration of the relatively brief 
period of 7 mins. 31 sees. Even a six-minute eclipse, such 
as the Sumatra eclipse of May 18, 1901 and its repetition 
eighteen years later in Brazil and Africa, is considered by 
astronomers unusually long. 

The discussion above regarding the conditions of a solar 
eclipse has been made with reference to the umbra of the 
moon's shadow. An observer on the earth situated within 
the confines of this shadow would find the sun's light totally 
obscured. Inside the penumbra of the moon's shadow, the 
sun is not entirely covered up by the moon and a partial 
solar eclipse will result. The diameter of the penumbra 
intercepted by the earth is readily determined. Since the 
angular diameters of sun and moon are nearly equal, the 
diameter of the penumbra measured at right angles to the 
line joining sun and moon is about twice the diameter of the 
moon, or roughly 4400 miles. Consequently, for a distance 
of 2 200 miles along the earth's surface on either side of the 
moon's umbra a partial eclipse may be visible. And by a 
process of reasoning similar to that followed out in the case 
of total eclipses, it is readily perceived that in latitudes away 
from the equator and when the eclipse is at sunrise or sunset, 
the distance along the earth's surface when a partial solar 
eclipse is visible may be increased to 3000 miles from the 
central line of totality. 

Again, since the sun and moon's angular diameters are 
about equal, and the moon moves the extent of her diam- 
eter in an hour, it will take about one hour from the first 


beginning of an eclipse, either solar or lunar, to the beginning 
of totality, or, as the astronomer expresses it, from first to 
second contact. Similarly an hour will elapse between third 
and fourth contacts, or between the ending of totality and 
the end of the eclipse. 

FIG. 2 

For a partial eclipse of the sun to occur, it is necessary for 
the moon at M to encroach on the shadow-cone ACDB en- 
veloping sun and earth. The angle between centers of sun 
and moon under these conditions is readily determined. 
The angle SEM in Fig. 2 is the sum of three angles, SEA, 
AEF and FEM. SEA and FEM are the angular semi-diam- 
eters of sun and moon respectively, while the angle AEF is 
equal CFE CAR. But CAR is the angle subtended by the 
earth's radius at the distance of the sun, and is accordingly 
the sun's horizontal parallax. If S and s represent the angu- 
lar semi-diameter of sun and moon, and P and p the hori- 
zontal parallaxes of sun and moon respectively, then SEM 
is equal to S + s P + p. 

For the solar eclipse to be total, the moon must be en- 
tirely within the shadow at m, and for these conditions the 
angle SEm is found to be S s P + p. 

For an eclipse of the moon, with moon at N, the angle be- 
tween the moon and the center of the earth's shadow, the 
angle NET is equal NEH + HET = NEH + CHE HTE. 
But SEA = EAT + HTE, hence NET = s+p S + P. 

Since the centers of the sun and earth are always in the 
ecliptic, the angles just determined give the amount in 
angle that the moon may be distant from the ecliptic at the 
time of new or full moon in order that there may be an 


eclipse of sun or moon. Angular distances measured at 
right angles to the ecliptic are called celestial latitudes, and 
hence in order that a partial eclipse may take place, the 
latitude of the moon must be less than 


p + s + (S P) for a solar eclipse, 
p + s (S P) for a lunar eclipse. 

Since S has a mean value of 32', while P is never greater 
than 9", it is evident that a solar eclipse can take place 
with the moon at a greater angular distance from the ecliptic 
than is possible for an eclipse of the moon. This is patent 
by referring to Fig. 2 and noting that a section across the 
shadow-cone enveloping sun and earth is greater through 
the moon at new moon, i.e., for a solar eclipse/than at full 
moon, or for a lunar eclipse. 

For the conditions under which total eclipses may occur, 
we find that the celestial latitudes of the moon at the time 
of conjunction of sun and moon are (by changing + s to s 
in the above formula) 

p s + (S P) for total solar eclipse, 
and p s + (S P) for total lunar eclipse. 

On account of the varying distances of sun and moon 
from the earth, the values of the semi-diameters and 
parallaxes of sun and moon are changing proportionally. 
In the table below the values are taken from the American 

Greatest Least Mean 

Semi-diameter of sun =5 16' 18" 15' 46" is'5o".6 

Semi-diameter of moon = s 16' 46" 14' 43" 15' 32". 6 

Horizontal parallax of sun = P 8".Q 8".y 8".8 

Horizontal parallax of moon = p 61' 28" 53' 55" 57' 2".7 

Inclination of moon's orbit = i 5 19' 4 57' 5 8' 43" 

An eclipse, either of sun or moon, might be predicted by 
finding from the above formulas the moon's celestial lati- 
tude. For purposes of computation it is more convenient 
to find the angular distance that the sun or moon may be 
from the moon's node, but measured in the plane of the 
ecliptic. These angles, being celestial longitudes, deter- 


mined at the time of new moon give what is known as the 
solar ecliptic limit, and at full moon furnish the lunar eclip- 
tic limit. In Fig. 3, NS is the ecliptic, MN the moon's path, 
N the moon's ascending node. M is the moon, and the circle 
at S, for a solar eclipse, represents a section of the shadow 
cone at M (in Fig. 2), and, for a lunar eclipse, a section 
through N (in Fig. 2), The angle SNM represents the 
angle, /, of inclination of the moon's orbit to the plane of 
the ecliptic. 


FIG. 3 

For a solar eclipse, the angular distance SM is known 
from above as equal to p + s + S -P. The angle i is known, 
and if the angle at S is assumed as a right angle the tri- 
angle SNM can be solved, and the distance SN determined. 
An approximate method is all that is required, and since 
the mean value of i is nearly one-eleventh of a radian, hence 
SN, the ecliptic limit is approximately eleven times SM. 
But as the value of SM depends on the semi-diameters 
and parallaxes of sun and moon, and as we have seen these 
quantities continually are varying, the value of the ecliptic 
limit must also vary proportionally, and also because the 
angle i changes likewise. Hence it is necessary to dis- 
tinguish between the maximum and minimum values of the 
ecliptic limit, for solar and lunar eclipses and for total and 
partial eclipses. Using the appropriate maximum and min- 


imum values, it is easy to find the following values of 
eclipse limits. 

Major Minor 

Solar eclipse 18 31' 15 21' 

Lunar eclipse 12 15' 9 30' 

Total solar eclipse 11 50' 9 55' 

Total lunar eclipse 6 o' 3 45' 

As a simple exercise in mathematical astronomy these 
quantities might be used for predicting eclipses. All of the 
additional information is readily obtainable from one of the 
government publications like the American Ephemeris and 
Nautical Almanac. As already stated, an eclipse of the sun 
can take place only at new moon. For this instant of time 
look up in the Ephemeris the longitude of the sun and that 
of the moon's node. The difference is the angular distance 
of the sun from the node. If this longitude is greater than 
the major ecliptic limit of i83i', an eclipse of the sun 
cannot possibly take place; if the quantity is less than the 
minor limit of i52i', an eclipse must certainly take place. 
If the difference in longitudes is less than i83i', but 
greater than 15 21', it is impossible to state whether an 
eclipse will take place or not, and recourse must then be had 
to a calculation of the ecliptic limit using the particular 
values for the semi-diameters, parallaxes and inclination for 
the day in question at the time of new moon. The condi- 
tions for a total eclipse are found by using the major and 
minor ecliptic limits of nSo / and 9 5 5'. Similarly, for an 
eclipse of the moon, which occurs only at full moon, by 
utilizing the values of the lunar ecliptic limits, it is possible 
to predict whether an eclipse of the moon will happen or not. 

In practice, however, in predicting the number of eclipses 
to occur in any one year, the eclipse limits defined above 
are not used. It is much simpler to use the Saros to bring 
forward the data of the Ephemeris of 18 years before. All 
of the various series of eclipses, lunar and solar, that are tak- 
ing place at the present time are known to the astronomer. 
When no great accuracy is needed the Saros provides a sim- 
ple method. The information so obtained should be checked 
by the magnificent book by Oppolzer, Canon der Finsternisse 


where are given the elements of all of the eclipses (8000 
solar and 5200 lunar) between the dates 1208 B.C. and 2162 
A.D. Various lists of total solar eclipses have been pub- 
lished. (See for instance. Encyclopaedia Britannica, article 
on Eclipses}. 

Referring now to the lengths of the different kinds of 
months on p. 40, we recall that 19 returns of the sun 
to the moon's node are equal to 6585.7806 days, while 
223 synodic months consume the slightly different amount 
f 6585.3211 days. If therefore a new moon fell exactly 
at the node, then after 18 years n days, the new moon 
would take place before the node was reached. The 
difference of the two quantities above is 0.4595 days, in 
which time the sun moves 28'. Accordingly, at each suc- 
ceeding Saros the new moon is found 28' farther and farther 
west of the node. It is now possible to trace the total 
number of eclipses and the progressive changes in them as 
they pass through the Saros. According to the value of 
the ecliptic limit already found, we know that if the new 
moon happens within 18 of the node, an eclipse of the sun 
may take place. If the node is the ascending one, the con- 
ditions will be as represented in Figs. 3 and 2, the moon 
being north of the shadow-cone. An eclipse of the sun 
will accordingly be visible in high northern latitudes on the 
earth. After 18 years n days the conditions will be nearly 
identical, but the new moon will take place 28' nearer the 
node, and as a result, the eclipse will be visible on the earth 
a little farther south of the preceding position. With each 
succeeding return, the new moon moves nearer and nearer 
the node and the eclipse track shifts farther and farther 
south on the earth. When the moon is within about 11 
of the node the solar eclipse becomes central, and the eclipse 
may be total or annular. As before, the central eclipse 
track will first touch only high northern latitudes, each 
succeeding eclipse moving farther and farther to the south. 
Total or annular eclipses will now take place at each 
return of the Saros until the new moon takes place 
11 west of the node, when the central eclipse track 
passes off the earth at its south pole, and a series of partial 


eclipses ensues until the moon is 18 west of the node, when 
these particular eclipses cease. In such a series there are 
from 68 to 75 solar eclipses, depending on conditions, extend- 
ing over some 1200 years. In each series there are approxi- 
mately 25 partial eclipses and 45 central eclipses. The 
numbers vary for different series of eclipses. Of the central 
eclipses, total eclipses follow total eclipses with about the 
same duration of totality, and annular eclipses follow an- 
nular. If the eclipses had taken place at the descending 
node of the moon's orbit instead of at the ascending node, 
eclipses would have come on the earth at its south pole and 
gradually moved north, going off the earth at the north pole, 
as shown in the illustration facing this page. 

On account of the ecliptic limits for lunar eclipses being 
smaller in value than for solar eclipses, there will be fewer 
repetitions in a lunar series, there being 48 or 49 altogether. 
Of this number there will be 22 or 23 total, with 13 or 14 
partial eclipses, both before and after the total eclipses. 
The interval for a series of lunar eclipses consumes about 
870 years. 

There is one other very important relation brought out 
by the data on page 41. It is there found that 239 returns 
of the moon to the line of apsides, or to perigee, amount as 
a total to 6585.5374 days. For predicting the nature of the 
eclipse that will be found at the next return this quantity 
is almost as valuable as the Saros itself. At the end of 223 
lunations the moon not only returns very closely to its origi- 
nal position with respect to the sun and the node, but also 
with respect to the line of apsides, with the necessary result 
that the distance from earth to moon is very closely re- 
peated. This fact in turn brings two consequences; firstly 
that the duration of the eclipse, or at least that part of it 
which depends on the moon's distance, is altered but little; 
but secondly, that the perturbations of the moon's orbit 
which otherwise might have displaced by several hours the 
time of the eclipse will now be repeated almost as they were 
before, and with no consequent relative effect on the time 
of the eclipse. 












< 2 

M +H Q 

B S ^ 

w a 









IN THE last chapter it was shown that the times of 
eclipses can be foretold with moderate accuracy by 
means of the Saros. Equally important with the actual 
prediction of the times is the fact that the circumstances at- 
tending the eclipses following each other in the Saros will be 
very closely repeated. A large partial eclipse of the sun will 
be followed by a large partial eclipse of the sun, an annular 
eclipse by an annular eclipse, a total eclipse of short dura,- 
tion by a similar short eclipse. And so also with lunar 
eclipses. There remains still to explain the signification of 
the decimal portion of the time of the Saros, 6585.3211 days. 
If the sun had been on the meridian at the middle of the 
eclipse, the eclipse therefore occurring at noon, the next 
following eclipse will be repeated at a place 0.3211 of a 
revolution of the earth, or in other words yh. 42m. of longi- 
tude farther west. After the return of three Saroses, or 54 
years, the eclipse tracks will have gone almost around the 
world and will have returned again to nearly the same 
longitude. If the eclipses belong to a series that is taking 
place at the moon's ascending node, the later eclipse track 
will be found farther south than the earlier one, and if at 
the descending node, farther north than the eclipse fifty- 
four years earlier. 

Those who wish to amuse themselves by playing with 
figures may find other remarkable coincidences by experi- 
menting with the lengths of the various years and months 
on p. 41. Newcomb has found a very interesting period 
at the end of 358 lunations. 

358 synodic months = 10571.95 days 
30.5 eclipse years = 10571.91 days 



This period amounts to 29 Julian years, less 20.3 days. But 
358 lunations are equal in length to 383.673 anomalistic 
months, and hence, as explained in the preceding chapter, 
the eclipses which follow each other with this period will 
differ greatly in their characteristics, since the conjunctions 
between sun and moon take place at different positions with 
respect to perigee. Also, such eclipses must follow each 
other alternately at ascending and descending nodes. Three 
such periods, however, will equal 1169.019 anomalistic 
months, and the time of conjunction has accordingly moved 
very close to perigee (if it had previously been found there) 
and a total eclipse will thus be followed by a total eclipse. 
Three of these periods equal 87 years less 61 days, and 18 
periods equal 521 years, actually within a day or two. Thus 
with the Saros of 18 years, and with the addition of the 
29-year, the 87-year and the 52i-year periods, the astrono- 
mer is enabled to engage in predicting at long range. Take 
for instance the first eclipse verified with certainty, the 
Nineveh eclipse of June 15, 763 B.C. (763 B.C. may be ex- 
pressed also as the year 762). Applying the 52i-year 
period we have the years 242 B.C. and A.D. 280, 801, 1322 
and 1843. These eclipses each fell on June 15, 0. S., or by 
the Julian calendar. The date of June 15, 1843, O. S., is the 
same as June 27, 1843, by the Gregorian or present-day 
calendar. By means of the 29-year period we obtain the 
eclipses of June 6, 1872, and May 18, 1901. (The eclipses 
of 1843 an d *872 were not total while that of 1901 was the 
six-minute total eclipse observed in Sumatra.) 

Facing page 56 will be seen a map of the tracks of all of 
the eclipses, total and annular, taking place between the 
years 1919 and 1940. The map is from Oppolzer's Canon 
der Finsternisse. 

In the following table are given one hundred years of 
total eclipses of the sun. The article on " Eclipses " in Ency- 
clopaedia Britannica, i4th edition, does not contain the 
eclipses of 1944, 1945, 1950, and 1968. As shown in the dia- 
gram facing page 52, the last two are unimportant eclipses 
which occur in the north polar regions. 






in minutes 

Where visible 

1875, April 6 




Indian Ocean, Siam, Pacific. 

1876, Sept. 17 




Pacific Ocean. 

1878, July 29 




Canada and the United States. 

1880, Jan. ii 



2 . I 

Pacific Ocean, California. 

1882, May 17 




Egypt, Central Asia, China. 

1883, May 6 




Pacific Ocean, Caroline Islands. 

1886, Aug. 29 




South America, Central Africa. 

1887, Aug. 19 




Northern Europe, Siberia, Japan. 

1889, Jan. i 



2. 2 

California, Oregon, Canada. 

1889, Dec. 22 




South America and Central Africa. 

1893, April 1 6 




Venezuela to West Africa. 

1894, Sept. 29 




East Africa, Indian Ocean. 

1896, Aug. Q 




North Europe, Siberia, Japan. 

1898, Jan. 22 



2 -3 

East Africa, India, China. 

1900, May 28 



2 . I 

United States, Spain, North Africa. 

1901, May 18 




Sumatra, Borneo. 

1904, Sept. 9 




Pacific Ocean. 

1905, Aug. 30 




Canada, Spain, North Africa. 

1907, Jan. 14 




Russia, Central Asia. 

1908, Jan. 3 




Pacific Ocean. 

1911, April 28 




Australia, Polynesia. 

1912, Oct. 10 




Colombia, Ecuador, Brazil. 

1914, Aug. 21 




Scandinavia, Russia, Asia Minor. 

1916, Feb. 3 




Pacific Ocean, Venezuela. 

1918, June 8 




The United States. 

1919, May 29 




Peru, Brazil, Central Africa. 

1922, Sept. 21 




East Africa, Australia. 

1923, Sept. 10 




California, Mexico, Central America. 

1925, Jan. 24 



2 4 

Northeastern United States. 

1926, Jan. 14 




East Africa, Sumatra, Philippines. 

1927, June 29 




England, Scandinavia. 

1929, May 9 




Sumatra, Siam, Philippines. 

1930, Oct. 21 




Pacific Ocean, Patagonia. 

1932, Aug. 31 




Canada, New England. 

1934, Feb. 14 




Borneo, Celebes, Caroline Islands. 

1936, June 19 




Greece to Central Asia and Japan. 

1937, June 8 




Pacific Ocean, Peru. 

1940, Oct. i 




Colombia, Brazil, South Africa. 

1941, Sept. 21 




Central Asia, China, Pacific Ocean. 

1943, Feb. 4 




China, Pacific Ocean, Alaska. 

1944, Jan. 25 




South America, West Africa. 

1945, July 9 



i .1 

Canada, Greenland, Scandinavia, 

1947, May 20 




South America, Africa. 

1948, Nov. i 




Central Africa, Madagascar. 

1950, Sept. 12 




Northeastern Siberia. 

1952, Feb. 25 




Africa, Persia, Central Asia. 

1954, June 30 




Canada, Scandinavia, Russia. 






in minutes 

Where visible 

I9S5 June 20 




Ceylon, Siam, Philippines. 

1958, Oct. 12 




Chile, Argentina. 

1959, Oct. 2 




Canaries, Central Africa. 

1961, Feb. 15 




France, Italy, Austria, Siberia. 

1962, Feb. 5 




New Guinea. 

1963, July 20 




Alaska, Canada, Maine. 

1965, May 30 




Pacific Ocean. 

1966, Nov. 12 




Bolivia, Argentina, Brazil. 

1968, Sept. 22 




Northern Siberia. 

1970, Mar. 7 




Mexico, Florida, Georgia. 

1972, July 10 




Northeast Asia, Northeast Amer- 

ica and Atlantic Ocean. 

1973, June 30 




South America, Africa. 

1974, June 20 




Southwest Australia and Indian 


From the table it will be seen that the total eclipses which 
have been or will be observed in the United States are: 1878, 
July 29; 1880, January n; 1889, January i; 1900, May 28; 
1918, June 8; 1923, September 10; 1925, January 24; 1932, 
August 31 ; 1963, July 20; 1970, March 7. So many eclipses 
in the past, so few in the coming years! In the future, the 
next eclipses to visit the United States will be those of 1945 
and 1954, both of which begin at sunrise about on the interna- 
tional boundary and their tracks will go northwards over inac- 
cessible spots in Canada. To observe either of these eclipses 
it will probably be necessary to travel to Scandinavia. The 
next eclipse to be observed scientifically in the United States 
is that of July 20, 1963, the path crossing Maine not very far 
northeast of the 1932 eclipse. The eclipse of March 7, 1970, 
will be visible in Florida and along the coast of Georgia. The 
next American eclipse after that will be on February 26, 1979, 
visible in northwest United States and in Canada across Hud- 
son Bay. What hardy astronomer will be willing to brave 
the probabilities of a blizzard and 40 F. below zero in order 
to add to scientific knowledge by observing this total eclipse? 

In the immediate future, the next total eclipse to be ob- 
served will be on June 19, 1936, with a duration of two and a 
half minutes. At sunrise the total eclipse begins near Greece, 
the path of totality later passing through Siberia and Japan. 

From Oppolzer's Canon der Finsternisse. 




The most favored sites for observation will be at the cities of 
Omsk and Tomsk on the Trans-Siberian Railroad. The as- 
tronomers from the United States will probably locate in 
Japan. The following eclipse is June 8, 1937, with a du- 
ration of totality of more than seven minutes. The track 
passes across the wide expanse of waters of the Pacific Ocean 
and touches land only near sunset in Peru. With the sun at 
a very low altitude it appears probable that no important sci- 
entific observations will be possible. The next eclipse, that 
of October i, 1940, will probably be extensively observed 
in South Africa. Evidently the coming generation of Ameri- 
can astronomers will be forced to make long trips away from 
home if they expect to continue the important discoveries 
made at eclipses. At the present time many eclipse problems, 
discussed in the final chapter, still await solution. It is con- 
fidently expected, however, that the coming generation will 
find methods of investigating chromosphere and corona with 
such success that observations at eclipses will no longer be 
necessary. When science shall have progressed to this ex- 
tent, future astronomers will be able to point to the long 
trips taken by their forefathers, sometimes half way round 
the globe, in order to observe a total eclipse for a few brief 
minutes of time. In the twenty-first century the great god 
of Efficiency may be worshipped even more profoundly than 
is the case in the early part of the twentieth century. Under 
such circumstances a total eclipse will still be viewed with 
interest as a fascinating phenomenon, which, however, will 
afford ample evidence of the crude methods of investigation 
of the astronomers a century before. 

If so few eclipses are visible in such a large country as the 
United States, how about those in a relatively small territory 
like the British Isles? The eclipse of June 29, 1927 was en- 
thusiastically observed in the northern portion of England 
in spite of heavy clouds; that of August n, 1999, grazes the 
south of Ireland and Land's End, and on August 12, 2026, 
the eclipse track may cross the southwestern tip of Ireland 
but will not be seen in Great Britain. 

Since the sixth century the following total eclipses have 
touched parts of the British Isles: 


594 July 23 1185 May i 

603 Aug. 12 1330 July 1 6 

639 Sept. 3 1424 June 26 

664 May i 1433 June 17 

878 Oct. 29 1598 Mar. 6 

885 June 15 1652 April 8 

1023 Jan. 24 1715 May 2 

1133 Aug. i 1724 May 22 

1140 Mar. 20 1927 June 29 

Eclipses have included London twice, Dublin twice, and 
Edinburgh five times. 

On the other hand, some parts of the globe are visited by 
eclipses in rapid succession. Spain witnessed the eclipses 
of 1842, 1860, 1870, 1900 and 1905. In the vicinity of the 
East Indies the eclipses of May 18, 1901, January 14, 1926, 
May 9, 1929 and February 14, 1934 have been observed, 
while the track of the eclipse of September 21, 1922, lay but 
a few hundred miles south. 

Assuming that the moon's shadow path intercepted by 
the earth averages 100 miles in width, it is easy to find that on 
the average a total eclipse of the sun will be visible in any one 
locality once every 360 years. A stay-at-home astronomer 
might have little opportunity during his lifetime to witness 
many such eclipses ! 

With a knowledge of the ecliptic limits found on page 50, 
it is possible to find the number of eclipses that will occur 
in any given time, say a year. If at the time of the new moon, 
the moon (and also the sun) is more than 18 31' from the 
moon's node there cannot possibly be an eclipse of the sun. 
Since the angle of 18 31' may be east or west of the node, 
there is thus a zone of 37, within which an eclipse of the 
sun may take place provided that a new moon also occurs 
somewhere within the zone. The minor ecliptic limit for a 
solar eclipse is 15 21', and similarly if new moon is found 
anywhere in the zone 30 42' in length an eclipse of the sun 
must happen, there being no possibility of the moon slipping 
by without intercepting the light from the sun as it comes in 
the direction towards the earth. For an eclipse of the moon, 
the ecliptic limits are smaller. It is only when full moon 
comes within 9 30" of the node that a lunar eclipse is certain 
to take place, though an eclipse may possibly happen if the 


limit is as great as 12 15'. In a synodic month the earth 
moves along the ecliptic to an average extent of 29 6'. But 
the moon's node is moving in the opposite direction by an 
amount of i 31' each month. Relative to the node, there- 
fore, new and full moon are found farther east each synodic 
month by the sum of 29 6' and i 31', or 30 37', and it is 
this motion relative to the node on which the circumstances 
of an eclipse depend. By an analogy one can see how easily 
the question of eclipses is thought out. Suppose a man walk- 
ing around a circular track of any diameter in which there are 
two mud holes four feet in width diametrically opposite each 
other. If he is taking a constant pace of three feet, it will be 
impossible for him to walk around the track without stepping 
at least once into each mud hole during each circuit. If 
when he comes to the muddy spot he happens to plant his 
foot near the middle he will have only one wet foot, but if 
the first foot gets in the hole near the edge, the second foot 
may be planted in the mud near the other side of the puddle. 
If the mud hole were only two feet in width, his standard 
pace of three feet might carry him across it without either 
foot getting into the mud. And so it is with eclipses of sun 
and moon. The moon's step is 30 37' in angular length, 
the danger zone for a solar eclipse is 30 42 ', or a quantity 
which is larger than the moon's monthly motion with re- 
spect to the node. There must, therefore, be at least one 
solar eclipse at each nodal point, or two such eclipses with- 
out fail during the course of each and every year. On the 
other hand, the danger zone for lunar eclipses being much 
smaller, it is possible that the full moon may take place so 
far from the node each time that there will not be a single 
lunar eclipse during the year. There can, moreover, be only 
one eclipse of the moon at each node, but never two. 
Eclipses do not occur in the same months each year, but 
follow each other at the beginning and end of the " eclipse 
year " of 346.6 days. This being 18 days less than the calen- 
dar year it is possible for there to be three eclipses of the 
moon within the calendar year, if the node is passed early in 
January, or late in December (which comes to the same 
thing). Two solar eclipses may possibly take place at each 


node, and a fifth one within the calendar year. If the new 
moon take place near the node bringing an eclipse of the sun, 
the preceding or following full moon may be too far from the 
node to cause an eclipse of the moon. Under these conditions 
there will be only two eclipses during the year, each of the sun, 
and as these occur near the node, they will be central eclipses, 
namely, either total or annular. This frequently happens, as, 
for instance, in the years 1922, 1926, 1929, 1933, 1940, 1944, 
1951 and 1962. If there are two solar eclipses at each node, 
a full moon must take place near the node and a total eclipse 
of the moon will result. There is thus the possibility of two 
partial solar and one total lunar eclipse at each node, or a 
total of two lunar and four solar eclipses during the year. 
If the node is passed about the middle of January, there may 
be a fifth solar eclipse within the calendar year, making a 
maximum of seven eclipses in a single year. There may 
also be three eclipses of the moon and four eclipses of the 
sun within the year, an event which will happen next in 1982, 
the three lunar eclipses each being total, and the four solar, 
partial. In 1 93 5 there will be two lunar and five solar eclipses. 
This combination occurs but seldom, and will not happen 
again until the year 2485. The maximum number of eclipses 
falling within a calendar year is thus seven and the minimum 

According to Oppolzer, there are 237.5 solar eclipses on 
the average in a century. Of these 83.8 are partial, 77.3 
annular, 10.5 annular and total (the vertex of the moon's 
shadow just reaching the earth) and 65.9 total eclipses. 
Twenty-seven percent of all solar eclipses are total. On the 
average, therefore, two total eclipses of the sun are visible 
every three years, but fully half of these are inaccessible for 
the reason that the eclipse tracks are in high northern or 
southern latitudes, or pass entirely across the oceans, or lie 
in localities where the probability of good weather is un- 
promising. Consequently, on the average, about once in 
every three years a total eclipse somewhere on the earth is 
available for observation. As the duration of totality aver- 
ages less than three minutes, we thus see that an eclipse 
observer may have an average of a minute per year for 



H >H 

H 4J 

oo~ PQ 




scientific work, provided he has sufficient financial backing 
to permit him to travel long distances for every available 
eclipse. The author has had the great good fortune of wit- 
nessing nine total eclipses in 1900, 1901, 1905, 1918, 1923, 
1925, 192 7, 1930 (Oct. 21) and 1932. In 1936 he hopes to be 
able to observe his tenth eclipse when he will have accom- 
plished the rare feat of observing three total eclipses in the 
same series separated by the Saros interval of 18 years n 

The Lick Observatory has set up an enviable reputation 
for itself through the observation of fifteen total eclipses in 
the following years: 1889 (Jan. i and Dec. 22), 1893, 1896, 
1898, 1900, 1901, 1905 (three parties in Labrador, Spain and 
Egypt)? i98, 1914? 1918, 1922, 1923, 1930 (April 28) and 
1932. In the past there have been but three complete failures 
through clouds, in 1896, 1914 and 1923. Excellent observa- 
tions have been secured by the British Joint Permanent 
Eclipse Committee under the auspices of the Royal Society 
and the Royal Astronomical Society. Through the coopera- 
tion of astronomers with eclipse experience working under 
the auspices of the Commission on Eclipses of the Inter- 
national Astronomical Union, problems are now carefully 
studied before each eclipse to the end that there may be little 
wasted effort. 

Taking into account the earth as a whole, it is evident that 
there are more solar eclipses than lunar, in the approximate 
ratio of four to three. But for eclipses visible from any one 
locality, London or New York or Timbuctoo, the question is 
an entirely different one. A partial solar eclipse at best may 
be seen from a limited portion of the globe, while a total 
eclipse is observable only in a very restricted path. An 
eclipse of the moon, being caused by its passage into the 
shadow of the earth, must be visible to every locality on the 
earth where the moon itself is visible. At any one instant 
of time the moon is above the horizon to half of the world, 
and since the eclipse of the moon, from beginning to ending 
of the. partial eclipse, lasts for some hours it will be plainly 
visible to over half the globe, unless clouds interfere. There 
are probably very few people living who love Nature, who 


have not witnessed at least one total eclipse of the moon, but 
probably not more than one in a thousand of the world's 
inhabitants has caught a glimpse of the matchless glory of 
the corona, seen only during the fleeting moments of a 
total eclipse of the sun. 


The seemingly uncanny and almost miraculous power of 
the astronomer to predict the coming of an eclipse hun- 
dreds of years in advance has always possessed a powerful 
fascination for the uninitiated. By means of the Saros 
and of the methods outlined in the foregoing chapter, it is 
readily possible to foretell the happening of an eclipse of the 
sun or moon and the general circumstances surrounding these 
events, provided no very great accuracy of time or place is 
required. To know the exact times for any given locality 
more accurate methods must be used. The problem of 
eclipse calculations resolves itself into three stages. First, 
the accuracy of the whole problem depends mainly on the 
completeness of our knowledge regarding the motions of 
the moon furnished by the lunar theory, and also upon the 
motion of the earth about the sun which gives the apparent 
motion of the sun about the earth, the latter theory being 
comparatively simple. From the motions of sun and moon 
thus determined, it is necessary to compute their positions at 
equidistant intervals, and as seen from a standard place, 
namely, the center of the earth. Second, from; these posi- 
tions it is necessary to compute certain " elements " on 
which eclipses depend; and third, for a given latitude and 
longitude of any place on the earth's surface, to calculate 
the times, etc., of the eclipse. The first part of the problem 
is taken care of by the various governmental publications 
which appear some three years in advance. By this means 
it is readily possible to find the exact positions of sun 
and moon at any given instant of time. Elements of eclipses 
needed for the second step of the process have been calcu- 
lated, from the earliest historic times up to the twenty- 
second century, by Oppolzer and Simon Newcomb. The 


third step in computing the details of the eclipses for a given 
year is now done by the American Ephemeris alone, but the 
results are published by the other nautical almanacs, thus 
eliminating duplication of efforts. For the 1932 eclipse 
Oppolzer's chart was in error by 200 miles in giving the 
eclipse path; his elements however furnished the position 
of the central line within an error of 10 or 15 miles and the 
times to an accuracy of about one minute. The American 
Ephemeris aims at an accuracy of one-quarter mile in posi- 
tion and two seconds for the times of contacts. 

The calculation of an eclipse of the moon is one of great 
simplicity. The time at which the moon passes into the 
shadow cast by the earth is the same absolute instant of 
time no matter where the observer is on the earth. Even 
the inhabitant of Mars (who may or who may not exist), 
who observes the earth for signs of life with the gigantic 
telescope which his supposed advanced civilization must 
have furnished him, would see the eclipse of the moon at 
identically the same absolute second of time as that noted 
by the Astronomer Royal at Greenwich. The local time of 
the eclipse recorded by the Martian clock would not be 
Greenwich mean time, but that is not part of the prob- 
lem. Eclipses of Jupiter's satellites, caused by the passing 
of these moons into the shadows cast by the planet are seen 
at the same instant of time by all observers on the earth, 
provided of course that the terrestrial observers are on the 
side of the earth turned towards Jupiter. If the various 
observers take the value of the Greenwich mean time of 
such an eclipse (which can be obtained from the nautical 
almanacs) and compare this with the local mean time of the 
individual observation, the difference in times will give the 
longitude of the observer from Greenwich. This is a method 
much in vogue for the determination of longitude when no 
great accuracy is necessary, for such an eclipse is a gradual 
phenomenon and not one which comes with great sudden- 
ness. The Greenwich mean times of the phases of an eclipse 
of the moon are the same for every inhabitant of the earth, 
and consequently, in each of the annual nautical almanacs 
it is possible to publish the Greenwich times of the various 


portions of the eclipse, such as the beginning of the partial 
eclipse, the beginning of the total eclipse, etc. If the person 
observing the lunar eclipse lives in Western Europe, in 
Great Britain, France, Spain, Portugal, Holland, Belgium 
or any of the other countries that now keep Greenwich time 
as standard, the times of the eclipse will be that recorded 
by their clocks, if these keep correct time. If the observer 
lives in New York or Washington or anywhere in United 
States or Canada where Eastern Standard time is used, the 
times of the eclipse will be obtained by subtracting 5 h., o 
m., o s. from the Greenwich time, while if he lives in San 
Francisco it will be necessary to subtract 8 h. from the 
Greenwich time. 

The method of calculating the times of a lunar eclipse, 
as given in the ephemerides, is a beautifully attractive prob- 
lem which may be solved quite simply by anyone even 
though he is not an expert computer, and this may be ac- 
complished either by graphical methods or by logarithmic 
calculation. Since the earth has an atmosphere, the outline 
of its shadow is not sharp and well-defined, but hazy when 
examined by a telescope. On account of the indefinite out- 
line of the earth's shadow, its theoretical diameter must be 
augmented. Different computers utilize different values for 
this increase, varying from one-fiftieth to one-seventy-fifth 
of the theoretical value of the diameter of the earth's 
shadow. Accordingly, a total lunar eclipse is not a sudden 
phenomenon like an eclipse of the sun and there is, there- 
fore, no need to employ very refined methods of computa- 
tion, or to carry the calculations to fractions of a second 
of time. The American Ephemeris gives the times to the 
decimals of a minute, an accuracy sufficiently great for 
the purpose. The method of calculating a lunar eclipse is 
given in Chauvenet, Spherical and Practical Astronomy, 
Vol. i, p. 589. For graphical methods of determining the 
times of eclipses of moon and sun, and also of occultations, 
see a series of articles by Rev. W. F. Rigge in Popular As- 
tronomy, volume III. 

As just stated, an eclipse of the moon does not permit of 
an exact calculation owing to the ill-defined nature of the 


earth's shadow caused by its atmosphere. Due to the re- 
fraction of light by the earth's atmosphere, the sun's light, 
or rather the rays at the red end of the spectrum that are 
not absorbed by the atmosphere, reach the moon, even 
at the middle of totality when the earth is directly between 
the moon and the sun. If the " man in the moon " were 
to view the phenomenon that would form, to him, an eclipse 
of the sun, he would see the earth surrounded by a great 
ring of light, its own illuminated atmosphere. If the moon 
had even a rare atmosphere, there would, for this same 
reason, be a ring of light around the moon as it en- 
croached upon the face of the sun at the time of a 
solar eclipse. No such light is visible, which is a sure 
proof that any atmosphere the moon may possess must be 
rarer than can be made by the best vacuum pump ever con- 
structed by man. The coming of totality at the time of an 
eclipse of the sun is a sudden phenomenon and as a conse- 
quence the approximate methods utilized in calculating a 
lunar eclipse are not sufficiently precise. The best method 
to follow in determining the times of a total eclipse is that 
of Chauvenet, Spherical and Practical Astronomy, i, 436. 
Another excellent guide is Buchanan, Theory of Eclipses. 

The calculation of a solar eclipse, in fact, cannot be 
treated in the simple manner of that of a lunar eclipse owing 
to the large size of the moon's parallax. In other words, the 
moon is so comparatively near the earth that it is projected 
on the face of the sun differently for every separate place on 
the earth's surface. As a result, the times of beginning and 
ending of the partial and total phases of the eclipse are dif- 
ferent at every station. 

The accuracy with which the times of solar eclipses can 
be predicted depends on the reliability of the work of the 
astronomers of all ages, and on the manner in which the 
torch has been passed on from one generation to the next. 
The chief cause of concern is found in the motions of our 
unruly neighbor, the moon. The position of the moon is 
furnished from observations of the times of contact of the 
limbs of the sun and moon, four different contacts being 
recognized. First contact is the instant that the moon be- 


gins to creep on the face of the sun, the eclipse beginning on 
the western edge of the sun. Second contact is signaled 
by the beginning of totality, third contact by the ending of 
totality and fourth contact is the ending of the eclipse, the 
moon passing off the face of the sun, the last contact being 
found on its eastern edge. The four contacts are generally 
observed visually by a pair of field glasses or by a telescope 
of moderate power. The time of first contact is difficult to 
observe with accuracy since nothing is to be seen at the edge 
of the sun until the moon is actually projected on to the face 
of the sun, and first contact has already actually taken 
place. In other words, the observer is always too late in 
noting the time of first contact, the amount of tardiness 
depending on the size of the telescope, the state of the see- 
ing, but especially on the skill and experience of the ob- 
server, which in turn depend on the number of eclipses wit- 
nessed. Fourth contact is easier to observe than first since 
the moon can be followed in the telescope until it leaves the 
face of the sun. The beginning and ending of totality can be 
more accurately observed than the two other contacts, but 
they are subject to some uncertainties on account of irregu- 
larities in the profile of the moon at the points of contact. 
By means of photographs taken just before and just after 
totality, when the crescent of the sun is small and changing 
rapidly, the times of second and third contacts can be deter- 
mined with much greater degree of precision than is attain- 
able in visual work. It goes without saying that the error 
of the chronometer should be known as accurately as 
possible (probably obtained by wireless signals), that the 
observed times should be recorded on the chronograph, and 
that the latitude and longitude be known. 

The photographs for contacts will unquestionably be 
taken with the same instrument used during totality for the 
corona. It will be necessary, therefore, to utilize a different 
method of exposure and different technique from that em- 
ployed on the corona. The brilliancy of the crescent sun 
compared with the corona will be very great. To diminish 
the brightness of the crescent images it will be necessary to 
use a rapid exposing shutter. The best results photographi- 


cally will be secured by using slow, fine-grained plates, and 
they should be specially " backed " to prevent halation 
caused by reflection from the glass side of the plate. (See 
Chapter VII.) 

At the eclipse of 1905, totality came ahead 1 of its pre- 
dicted time, the beginning of totality being 17 seconds 
earlier, while the ending came 23 seconds earlier than the 
times predicted by the American Ephemeris. The middle 
of totality was thus 20 seconds ahead of that calculated, 
while the duration of totality was some six seconds less than 
was expected from the computations. The time predicted 
for the middle of totality by the British Nautical Almanac 
was identical with that furnished by the American Ephem- 
eris but the duration of the former was i s .7 less than that 
of the latter, while the duration calculated from the Con- 
naissance des Temps was five seconds greater than that of 
the American Ephemeris. All observers at the Spanish 
eclipse had their program of observation greatly interfered 
with by having the moon so far in advance of its predicted 
place. Before the eclipse of June, 1918, the observers of 
the U. S. Naval Observatory party in Oregon were furnished 
by the Washington authorities with a correction of 12.5 
seconds to be applied to the times of contacts calculated 
from the American Ephemeris. The observed times were 
about fourteen seconds ahead of those predicted by the 
Ephemeris. At the eclipse of 1922, the Lick Observatory 
party observed the beginning of totality some sixteen 
seconds earlier than the time predicted, while the end of 
totality came twenty seconds earlier than the Almanac 

Apparently, therefore, after having made due allowance 
for all known possible sources of error, the moon has strayed 
from the path mapped out for it by the mathematical astron- 
omers by an amount which is not inconsiderable. What 
is the cause of the moon being away from its predicted 

The first of the great modern authorities dealing with the 

1 Lick Observatory Bulletin, 4, 118, 1905. 


motion of the moon was Hansen, who, in 1857, published 
his Tables de la Lune. By the help of these tables the un- 
explained fluctuations in the moon's motion were reduced 
to a small fraction of their former amount. Hansen de- 
rived the inequalities in the moon's motion largely from the 
gravitational theory, but in order to satisfy the observa- 
tions made with great care at Greenwich between 1750 and 
1850, it was necessary for him to apply two empirical terms 
supposed to come from Venus, the larger of the two having 
a long period of 239 years. 

The next advance was made by Simon Newcomb, who 
published his Researches on the Motion of the Moon in 
1878. In addition to the observational material utilized by 
Hansen, Newcomb discussed all of the lunar eclipses re- 
corded by Ptolemy in the Almagest and in addition a large 
number of mediaeval lunar eclipses. He discussed a few 
ancient solar eclipses, but excluded them from his calcula- 
tions on account of their unreliability. The chief value of 
Newcomb 's great work lay in collecting and discussing 
observations of occultatioris and solar eclipses in the cen- 
tury and a quarter previous to 1750. Newcomb had 
thus 250 years of observations to discuss in place of 
the hundred years of observations available to Hansen. 
The whole material was utilized to secure the value of the 
moon's motion and acceleration, and clearly showed the 
presence of an unexplained term of long period. Newcomb 
did not deduce the period of this term from the observations 
themselves but assumed it identical with one of Hansen 's 
terms (which is now known to be faulty). " It is no small 
tribute to the thoroughness of Newcomb 's work that while 
the observations from 1750 onwards have been thoroughly 
examined by Cowell, Radau, and Brown, the reductions of 
the observations from 1621 to 1747 have never been revised 
by anyone but Newcomb himself." x 

The third great investigation of the lunar theory is found 
in E. W. Brown's Tables of the Moon y s Motion, published 
in 1920. These tables are more complete than those of 
Hansen and Newcomb, and take account of every term of 

1 Fotheringham, Monthly Notices, R. A. S., 80, 289, 1920. 


appreciable significance. In addition to these three great 
monumental works on the Theory of the Moon by Hansen, 
Newcomb and Brown, exhaustive investigations have been 
made by many competent authorities. After making allow- 
ance for the gravitational attraction of every conceivable 
form of disturbance, it is unmistakable that the moon de- 
Darts from her theoretical place in a very irregular manner. 
An error in the assumed value of the acceleration of the 
moon's mean longitude, even though this error is a very small 
3ne, will cause errors in the predicted place of the moon which 
increase in size for the reason that the errors depend on the 
square of the elapsed time. 

Tables of the Moon are used for predicting the place of 
the moon in the different nautical almanacs which appear 
some two or three years in advance. For these predictions 
It is desirable to keep the theory of the moon as free as 
Dossible from arbitrary empirical terms so that the theory 
nay not be cluttered up by too many additions. When ob- 
servations are secured the results can then be directly com- 
pared with theory. Between the theoretical position of the 
noon given by the almanac, and the observed place, there 
nay be a difference of several seconds of arc. 

For guiding the work of the eclipse astronomer it is now 
leemed necessary to know the times of beginning and ending 
)f a total solar eclipse within an error of a few seconds, 
rlence it has become customary in recent years to secure 
Tom Washington, or Greenwich, a month or more before the 
eclipse takes place, corrections to the almanac positions of 
;he moon. These corrections derived from observations in 
he three-year interval since the almanac was prepared per- 
nit increased accuracy in computing the eclipse. 

On account of the difficulty of seeing the moon when near 
he time of new moon, there is ordinarily a gap in the regular 
unar observations, whether these are made by occultations 
>r by meridian circle. An intense interest was aroused in 
he United States by the total eclipse of 1925 and hence a 
pecial attempt was made to utilize this interest in order to 
iecure as many observers as possible. From large numbers 
>f careful observations it was hoped that it might be pos- 



sible to detect any short-period deviations in the motion of 
the moon, as well as to compare the results obtained by dif- 
ferent methods of observation in order to see whether there 
are systematic errors of observation peculiar to any of them. 
At the time of the 1925 eclipse it was found that the dif- 
ferences in the observed and tabular positions of the moon 
had been nearly constant during the interval since the 
Ephemeris was published. For predicting the eclipse, it was 
necessary to apply a correction of + 7^.0 to the tabular 
mean longitude of the moon as given in the American 
Ephemeris and 0^.50 to the lunar declination. After ap- 
plying these changes to the moon's place, the observations of 
the beginning and ending of totality gave the following cor- 
rections to be applied to the predicted times of second contact 
and duration of totality, the observations coming from 
trained astronomers with every facility for obtaining accu- 
rate positions of their stations and accurate time signals on 
the day of the eclipse. 


Poughkeepsie, N. Y. 
Beacon, N. Y. 
New Haven, Conn. 
Middletown, Conn. 
Martha's Vineyard, Mass. 
Nan tucket, Mass. 
Dirigible, Los Angehs 


+ 3-M 
+ i 8 

4- () o 

+ 3 6 
+ 60 

+ 49 


o. s 8 

- 2 Q 
+ 08 

O O 


- 3 7 

- 0.9 

A summary of the results of the observations discussed by 
Brown 1 for the 1925 eclipse gives corrections to be added to 
the mean longitude of the moon 8X, and to the mean latitude 
S/3 as follows, where n represents the number of observations: 


6X 7 "oo 





+ o "38 ."10 


+ o "40 db "17 



o 14 .25 


+ o 34 dc . 26 

i i 

Greenwich, meridian circle 

o 35 dc .16 


85 rb .11 


Washington, g-inch transit 

- o 34 15 


o 60 .16 


Washington, 6-inch transit 

+ 70 IQ 


0.13 .18 


Cape, meridian circle 

+ o 42 d= 28 

J 3 

o 34 dr .46 



-f- O 40 dr 14 

+ 80 .12 

1 Astronomical Journal, 37, 9, 1926. 


The weighted mean of the results gives a correction to the 
moon's longitude and latitude of + 0.^28 ."06 and 
+ o."32 ."06, respectively. 

The comparison of the results made by the different 
methods is not without interest. The meridian observations 
differ rather widely. The occultations and eclipse results 
give corrections which agree within their probable errors. 
All the non-meridian observations agree in giving a positive 
correction to the lunar latitude and all the meridian results 
give a negative correction, and hence systematic differences 
of over i."o are shown. These discrepancies are partially ex- 
plained by an assumption originally made by Hansen that 
there is a difference between the center of mass and center of 
figure of the moon. Brown has shown that the northern half 
of the moon is probably the denser half. 

The ranges in the times of second contact and the duration 
of the eclipse, as exhibited in the table above, based on ob- 
servations of skilled astronomers, appear much larger than 
one would expect, but the observations of contacts during 
the eclipse of 1914 contained in Meddelande jran Lunds 
Astron. Obs., No. 20, show about the same differences. How- 
ever, it must not be forgotten that the outline of the moon 
departs from a perfect circle and that the time of the begin- 
ning and ending of a total solar eclipse depends very largely 
on the character of the lunar surface at the point of contact, 
the beginning of the total eclipse not taking place until the 
last Baily bead has disappeared. 

The marvelous accuracy in the prediction of eclipses shows 
the wonderful precision and perfection of the science of as- 
tronomy. An observation of an eclipse of the sun made three 
thousand years ago has an important bearing on the most 
recent refined researches on the motion of the moon ! 



ASTROPHYSICS, called the " new astronomy/ 7 has 
revealed in a remarkable manner through its 
discoveries during the past half century the won- 
derful ability and resourcefulness of the human brain, mak- 
ing it evident that man is gifted with almost infinite powers. 
From this earth of ours, which astronomy teaches us to be 
but an insignificant speck among the countless orbs of the 
universe, man has been able to reach out and ascertain the 
physical constitution of the sun, and to acquire this in- 
formation with almost the certainty of a chemist who could 
make a qualitative analysis provided that an average speci- 
men of the sun's matter could be furnished him. And 
across millions and millions of miles of space, the far dis- 
tant suns, the stars, shining by their own feeble light, are 
made to give up the secrets of their construction. Not 
only can we learn of the constitution of the sun and stars, 
but we can also ascertain their effective temperatures as 
well, and with information thus garnered we can arrange 
the stars in an orderly sequence, tracing their evolution 
from the swollen red stars of very minute density, the so- 
called " giants/ 7 through the successive stages of yellow 
and white to that of the stars of class B, and then on in 
the descending branch of development as the stars become 
cooler and more dense. The astronomer believes in evolu- 
tion and recent researches make evident that our gigantic 
sun is but a yellow " dwarf/ 7 well advanced towards the 
state of old age and final obscurity. It is by means of the 
spectroscope that such information is gathered, and by the 
spectroscope it has become possible to investigate motions, 
not athwart the sky as the older astronomy was able to do, 
but towards us or away from us in the line of sight, and 
to measure these motions in miles per second. The shift 



of the lines of the spectrum due to motion in the line of 
sight, which has been confirmed experimentally in the 
laboratory, has given rise to many interesting developments 
in astrophysics: the discovery of an entirely new class of 
bodies, spectroscopic binaries, the measurement of the 
axial rotation of the sun and Jupiter, as well as providing a 
magnificent confirmation of the meteoric constitution of 
Saturn's rings. And quite recently the greatest triumph 
of the spectroscope has been achieved by Adams and his 
co-workers at Mount Wilson Observatory being able to 
determine the distances of the stars and thus find their 
luminosities compared with that of the sun. Since its 
birth in 1859, when Kirchhoff discovered the principles of 
spectrum analysis, astrophysics has advanced by leaps and 
bounds. In no branch of astronomy has the new instru- 
ment of research shown its outstanding value quite as clearly 
as in the development of the subject of eclipses of the sun. 
For a clear understanding of the matter it will therefore 
be necessary to give a brief account of the history of the 
new astronomy. 

The scientific world owes much to John Kepler for hand- 
ing down to us the three great laws of planetary motion, but 
his activities extended beyond the realm of mathematical 
astronomy into the domain of physics. He was the first to 
show as he did in his " Dioptrics " that if a beam 
of sunlight be allowed to fall upon a prism in a certain way 
it passes out as a colored beam, giving the colors of the 
rainbow. And as the great Sir Isaac Newton took Kepler's 
Laws and used them for finding the law of gravitation, so 
likewise he extended the physical work of Kepler and by 
reason and experimentation greatly increased our knowledge. 
In fact, it is from Newton's labors that the science of spec- 
trum analysis virtually had its birth. 

In 1666, Newton, by allowing sunlight to pass through a 
hole in a shutter and to fall on a glass prism, found a colored 
ribbon of light, an impure spectrum. His description of the 
effect observed is so clear that the following is copied from 
Optics, Third edition, page 21." 

1 See also, The Spectroscope and its Work, Newall, 191 1. For greater details 
consult Chemistry of the Sun, Lockyer, 1887, and Handbuch der Spectroscopie, 
Kavser, Vol. i. 


" In a very dark chamber at a round hole about one- 
third part ot an inch broad made in the shut of a window 
I placed a glass prism, whereby the beam of the sun's light 
which came in at that hole might be refracted upwards 
toward the opposite wall of the chamber, and there form a 
coloured Image of the Sun. The axis of the prism (that is 
the line passing through the middle of the prism from one 
end of it to the other end parallel to the edge of the refract- 
ing angle) was in this and the following experiments per- 
pendicular to the incident rays. About this axis I turned 
the prism slowly, and saw the refracted light on the wall 
or coloured image of the sun first to descend, and then to 
ascend. Between the descent and ascent when the image 
seemed stationary, I stopped the prism and fix'd it in that 
posture, that it should be moved no more. For in that 
posture the refractions of the light at the two sides of the 
refracting angle, that is at the entrance of it, were equal to 
one another. So also in other experiments, as often as I 
would have the refractions on both sides of the prism to be 
equal to one another, I noted the place where the image of 
the sun formed by the refracted light stood still between its 
two contrary motions, in the common period of its progress 
and regress; and when the image fell upon that place, I 
made fast the prism. And in this posture, as the most con- 
venient, it is to be understood that all the prisms are placed 
in the following experiments, unless where some other pos- 
ture is described. The prism therefore being placed in this 
posture, I let the refracted light fall perpendicularly upon 
a sheet of white paper at the opposite wall of the chamber, 
and observed the figure and dimensions of the solar image 
formed on the paper by the light. This image was oblong 
and not oval, but terminated with two rectilinear and paral- 
lel sides, and two semicircular ends. On its sides it was 
bounded pretty distinctly, but on its ends very confusedly 
and indistinctly, the light there decaying and vanishing by 

Newton concluded that white light was made up of 
separate colored rays; by passing the light through the 
prism, the different rays suffered different amounts of re- 


fraction; they were in fact dispersed, and as a result the 
spectrum consisted of an infinite number of colored images 
of the round hole lying side by side. In the middle of the 
spectrum the different rays overlapped, and white light 
resulted, but the ends remained colored. Newton found 
that by the use of a slit " an inch or two long, and a tenth 
or a twentieth of an inch in width,' 7 the spectrum became 
purer; and he even tried a triangular hole, observing 
greater and greater purity as the vertex was approached. 
He proved conclusively that the colors came from the sun's 
light itself and not from the prisms, for he produced spectra 
with a variety of different prisms, and afterwards combined 
the prismatic colors together to make white light. He 
further showed that each ray of light consisted of a single 
color and possessed a certain definite refrangibility. 

Newton's books on optics (published in 1704) are mar- 
vels of clearness of exposition, and his fundamental ex- 
periments have come down to us almost unchanged. The 
significant result of Newton's work is that the pure colors, 
and not white light, occupy the primordial position of im- 
portance, since it is possible to form all conceivable colors in- 
cluding white light from a mixture of the pure colors. A 
determined color in the spectrum is defined through its index 
of refraction, and since this quantity continually varies, 
there is an infinite number of spectral colors. 

Newton did not see any of the dark lines of the solar 
spectrum, now known as Fraunhofer lines, though he used 
a narrow slit and should have been able to see them, but 
his prisms were poor (as he himself said). The authority 
of his great name discouraged further experimentation, and 
no advances were made for one hundred years. Unfortu- 
nately he made some mistakes, first in dividing the colors 
into seven, the quantitative number of such great attrac- 
tion and " perfection " to the early scientists, thereby per- 
haps preventing the discovery of the Fraunhofer lines, and 
second, in not seeing that various media dispersed differ- 
ently. As is well known, Newton said that the case of the 
refracting telescope was a deplorable one, and following 
his time these telescopes were made of enormous lengths, 


up to two hundred or even three hundred feet, in order to 
minimize the color difficulty. However, if Newton had 
been in a position to make use of different kinds of glass 
of crown and flint as Fraunhofer was later, it is altogether 
likely that many other discoveries of great importance 
would have been made by him. 

The long period of arrested progress was broken in 1802 
when Wollaston, making use of a slit, repeated Newton's 
experiments. He then found, 1 " The colours into which a 
beam of white light is separated by refraction appears 
to me to be neither seven, as they are usually seen in the 
rainbow, nor reducible by any means to three, as some 
persons have conceived; but that, by employing a very 
narrow pencil of light, four primary divisions of the 
prismatic spectrum may be seen with a degree of distinct- 
ness that, I believe, has not been described nor observed 
before. . . . The four colors are, red, yellowish-green, blue 
and violet." He then goes on to describe the dark markings 
that were seen in his spectrum, these being ill-defined in 
appearance and evidently taken as the natural divisions 
between the colors. It is readily apparent that the prisms 
used by Wollaston were likewise poor in quality, for his 
experiments were made in a manner that should have 
brought to view the thousands of dark lines in the solar 
spectrum. However, a very great advance was made over 
the work of Newton, for the spectrum of a candle flame 
and of an electric light were examined. Spectra were found 
which not only differed from each other in appearance but 
each of which was entirely unlike that furnished by the 
sun. The candle and the electric light each gave a spec- 
trum of bright lines, a discontinuous spectrum. The funda- 
mental importance of these experiments was not at the time 

The life of Joseph Fraunhofer 2 had an almost tragic 
beginning. At the age of fourteen he lived in a dilapidated 
house in an alley in Munich which tumbled down and 
buried its occupants in the ruins. The other residents were 

1 Phil. Trans., part i, 378, 1802. 

2 Clerke, History of Astronomy during the Nineteenth Century. 


killed, but the boy, who was an orphan, was dragged out. 
more dead than alive and seriously injured. The Elector 
Maximilian Joseph was a witness of the accident and to 
show his commiseration made him a present of eighteen 
ducats. Besides the purchase of books and of a glass- 
grinding machine the money permitted his release from ap- 
prenticeship to a looking-glass maker. Through study and 
toil and privation he increased his knowledge of the 
optician's art and at the age of nineteen entered the glass- 
making firm of Von Reichenbach and Utzschneider. He de- 
voted himself now with great avidity to a study of lenses 
for the purpose of improving the refracting telescope. After 
Newton's time, Dollond had discovered that it was possible 
to banish most of the color which so interfered with the 
action of the refracting telescope, by combining a lens of 
crown glass with one of flint. By means of many experi- 
ments with prisms of glass of different varieties, Fraunhofer 
was able to investigate the best combination of two lenses 
that would give the most perfect definition with freedom 
from the disturbing color. In 1817 there was finished the 
great " Dorpat refractor " with the then unprecedented 
aperture of nine and a half inches. This telescope in the 
skillful hands of Struve became one of the most famous 
telescopes ever in existence. To Fraunhofer the astronomi- 
cal world is also indebted for the first really serviceable 
heliometer, that at Konigsberg, an instrument which was to 
play an important part in extending our knowledge of 
the sidereal universe by permitting the measurement of 
stellar distances, or, as they are technically called, stellar 

In 1814, Fraunhofer not only extended Wollaston's work 
but introduced great improvements in the method of ob- 
serving. The slit was retained but placed at a great 
distance from the prism. Instead of allowing the refracted 
beam of light to fall on a screen he placed a small telescope 
directly behind the prism and by this means a magnified 
view of the spectrum was obtained. 

" Into a dark room, and through a vertical aperture in 
the window-shutter, about 15" broad and 36" high, I intro- 


duced the rays of the sun upon a prism of flint glass placed 
upon the theodolite; this instrument was 24 feet from the 
window, and the angle of the prism was nearly 60. The 
prism was placed before the object glass of the telescope, 
so that the angles of incidence and emergence were equal. 
In looking at this spectrum for the bright line which I had 
found in the spectrum of the artificial light, I discovered, 
instead of this line an infinite number of vertical lines of 
different thickness. These lines are darker than the rest of 
the spectrum, and some of them appear entirely black." * 

The interrelation of these lines and streaks appears to be 
the same no matter what refracting substance is employed 
so that, for instance, a particular band is found in each case 
only in the blue, another is found only in the red, and one 
can therefore learn to recognize a particular line in the spec- 
trum by noting its position with respect to the prominent 
lines. Fraunhofer observed that the strong lines did not 
mark the edges of the colors as Wollaston had supposed 
and further that the same color is found on both sides of the 
line, the colors grading by imperceptible degrees from one 
color to the next. Starting from the violet end of the spec- 
trum, the colors are given the following names: violet, in- 
digo, blue, green, yellow, orange and red. As Lockyer has 
pointed out, the first letters of these color names make the 
combination VIBGYOR, an aid to memory that the beginner 
may find useful. 

Fraunhofer constructed a map of the lines of the solar 
spectrum, measuring by means of the circle of the theo- 
dolite the accurate positions of over 350 lines, though 
he counted no less than 754. Starting at the red end of the 
spectrum he called the more prominent lines by the letters 
of the alphabet. Thus A, B and C denote lines in the red 
part of the spectrum, D the prominent double lines in the 
yellow part, which we now know to be due to sodium, while 
H and K in the violet are the very broad lines caused by cal- 
cium. In addition, the small letters of the alphabet were 
made use of, b, for instance, denoting a group of lines in the 
green due to the element magnesium. Not only are the 

1 Denkschriften der K. Akad. der Wissen. zu Miinchen, 5, 193, 1814. 


prominent lines of the solar spectrum to which Fraunhofer 
gave names called after their discoverer, but all dark lines 
in the spectrum whether prominent or faint are known as 
Fraunhofer lines. 

He plotted his lines not according to their wave-lengths 
as in the manner of all modern maps, but according to a 
rather arbitrary scale. Now that the solar spectrum has 
been more fully investigated with the perfected apparatus 
of the twentieth century, we can go back and gauge at its 
true value the worth of Fraunhofer's map. Thus we have 
the opinion of Hartmann of Potsdam to the effect that 
Fraunhofer made his map with the greatest degree of re- 
finement and secured a precision which has warranted spec- 
troscopists of the past hundred years placing in it the great 
confidence that they have felt. Fraunhofer's life and work 
is a splendid example of what one man by patient and care- 
ful work can accomplish for the cause of science. Any one 
who has ever looked into a spectroscope will realize what a 
colossal work the making of this map must have been. 

Since the lines and bands in the color image have only a 
very small width, it is evident that the apparatus must be 
most perfect to avoid all aberrations which could either ren- 
der the lines indistinct or entirely scatter them. The faces 
of the prism must therefore be perfectly plane; the glass 
to be used in such prisms should be entirely free from waves, 
streaks and striae ; and the greatest care should be exercised 
in their grinding and polishing. These and other considera- 
tions, such as, that the slit must be parallel to the edge of the 
prism, Fraunhofer found out by careful experimenting. 

Fraunhofer, however, was not content with this work. 
He wanted to know something of the origin of the lines, 
and he soon came to a conclusion on this point. It occurred 
to him that they might possibly be attributed to some illu- 
sion caused by the narrow aperture through which the light 
was admitted. We know that the shape of the slit has 
something to do with the forms of these dark spaces in the 
spectrum, but with their simple existence as spaces the slit 
has nothing to do, the mere shape of the lines being quite a 
trivial matter. To settle this question beyond doubt he 


changed the slit which he was using in order to ascertain 
if this would change his spectrum. He passed light through 
a small round hole of 15" in diameter, and allowed it to fall 
upon the prism placed in front of the objective of the theod- 
olite. It is clear that the color-image seen with the tele- 
scope can have only an inappreciable width, and therefore 
will form only a line; but in this narrow colored width no 
fine cross lines can be seen. In order to widen this narrow 
stretch of light into a band wide enough to see, Fraunhofer 
made use of a cylindrical lens, or a lens which is plane on 
one side and curved on the other resembling a portion of a 
cylinder of large diameter. The axis of the cylinder was 
placed parallel to the base of the prism and hence parallel 
to the line of light. Consequently, the width of the spec- 
trum would be changed without in any way altering its 
length. With this arrangement the lines were observed to 
be exactly the same as when the light comes through a long 
narrow opening. Hence the bands and lines in the solar 
spectrum are not caused by diffraction and interference by 
the light passing through the slit, nor are they produced by 
any peculiarities in the apparatus. There is, therefore, only 
one other possible cause for these lines, and that is that they 
somehow or other belong to the light given out by the sun. 
Hence the solar spectrum is not a continuous spectrum having 
light of all colors and all wave-lengths, but is, on the con- 
trary, a discontinuous one in which vibrations of certain 
lengths are missing from the sum-total which goes to make 
up white light. 

If the sun is remarkable for this discontinuous spectrum, 
what types of spectra do the other heavenly bodies show? 
If Fraunhofer could examine the light from a small round 
opening at a short distance from his instrument, there is no 
reason, if the light were sufficient, why he should not ex- 
amine the light from a round body made apparently small 
by the fact of its situation at a great distance. He ex- 
amined, therefore, the light from Venus directly without 
making the light pass through a small opening, and he found, 
after spreading out the light by means of the cylindrical 
lens ; that the same lines appeared in the light of Venus as 


appear in sunlight. Since, however, the light from the planet 
is very feeble in comparison with the light received from the 
sun, the intensity of the violet and red colors of its spectrum 
are very weak, and on this account even the stronger lines 
in both these colors are seen with difficulty, though in the 
other colors they are very easily distinguished. Fraunhofer 
was able to see D, E, b and F perfectly defined and he even 
recognized in the triplet b in the green, two lines, one weak 
and one strong, although he was unable to see that the 
stronger of these two lines was really a double line. This, 
of course, was due to the weakness of the spectrum, and for 
this same reason the other finer lines could not be distin- 
guished satisfactorily. By measuring the arcs DE and EF 
it was made certain that the light from Venus contained, as 
far as could be analyzed, just the same lines in its spectrum 
as did the light of the sun. 

With this same apparatus observations were made on the 
light of some fixed stars of first magnitude, but since the 
light of these stars is much weaker than that of Venus, it 
is natural that the brightness of the spectrum should be 
much less. In spite, however, of this comparative lack of 
brilliancy, Fraunhofer was able to recognize with certainty 
in the spectrum of Sirius three broad lines which appear to 
have no connection with those of sunlight: one of these is 
in the red, one in the green, and the other in the blue. In 
the spectra of other fixed stars of the first magnitude, lines 
were actually recognized by him, and it seemed certain that 
these spectra though very faint differed amongst themselves. 
The observations were made with a telescope of an aperture 
of about one inch, so that it was impossible for the distin- 
guished pioneer to do more than to point out the way. In 
such a manner as this was the science of stellar spectroscopy 

As we have seen, Fraunhofer examined the stellar spectra 
by allowing the light from the star to fall directly on the 
prism and after refraction to examine this light by the 
telescope. This same method of observation is still of the 
greatest scientific value and it is in this way that Harvard 
College Observatory has been able to accomplish such an 


enormous amount of sound research by methods involving 
the use of the spectroscope. The combination of prism and 
object glass is called an " objective prism " or, when used 
for photographing, the " prismatic camera." If we point 
such an instrument to any place in the sky, we can view by 
our eye, or photograph on a plate, the spectrum not of one 
star only, but of all the stars that are in the field of the 

It is not to be wondered at that a man who had thus 
brought sun, planet and star within the grasp of a new in- 
strument should not rest content with these observations. 

Fraunhofer next investigated at considerable length the 
spectra of artificial light. In an early part of his paper he 
states that, on examining the spectra of flames, he found 
that flames such as that of a lamp and candle, and, indeed, 
in general the light produced by the flame of a fire, exhibit 
between the red and yellow of the spectrum a clear and well 
marked line which occupies the same place in all the spectra. 
Returning to this subject later, he notes that in transmitting 
the light of a lamp through the same aperture employed for 
the examination of the solar spectrum, a line appears which 
corresponds exactly to the position of the D line in the solar 
spectrum. In fact, the resemblance to the solar D line is so 
close that both the bright artificial line and the solar D line 
are each a fine double line. This was the first step towards 
the true explanation of the dark lines in the solar spectrum; 
but it took many years before the true meaning was arrived 
at, mainly on account of the presence of this bright D pair 
in practically every flame and under all conceivable sets of 

According to Agnes M. Clerke, 1 " the ubiquity and con- 
spicuousness of the sodium-line long impeded progress. It 
was elicited by the combustion of a surprising variety of 
substances sulphur, alcohol, ivory, wood, paper; its per- 
sistent visibility suggesting the accomplishment of some uni- 
versal process of nature rather than the presence of one in- 
dividual kind of matter. But if spectrum analysis was to 
exist as a science at all, it could only be by attaining cer- 

1 History of Astronomy during the Nineteenth Century. 


tainty as to the unvarying association of one special sub- 
stance with each special quality of light. It appeared, in- 
deed, without fail where sodium was; but it also appeared 
where it might be thought only reasonable to conclude that 
sodium was not. Nor was it until thirty years later that 
William Swan, by pointing out the extreme delicacy of the 
spectral test, and the singularly wide dispersion of sodium, 
made it appear probable (but even then only probable) that 
the questionable yellow line was really due invariably to 
that substance. Common salt (chloride of sodium) is, in 
fact, the most diffusive of solids. It floats in the air; it flows 
with water; every grain of dust has its attendant particle; 
its absolute exclusion approaches the impossible. And 
w r ithal the light that it gives in burning is so intense and 
concentrated, that if a single grain be divided into 180 mil- 
lion parts, and one alone of such inconceivably minute frag- 
ments be present in a source of light, the spectroscope will 
show unmistakably its characteristic beam." 

The advance made by Fraunhofer in studying the re- 
fraction of light through prisms placed the new science on 
a very firm foundation. This remarkable progress, due to 
the very marked improvement in the quality of the glass 
used in the prisms, was rendered possible only by the em- 
ployment of a telescope of greatly increased defining power. 
The utilization of the telescope permitted advances in the 
study of diffraction of light through a narrow aperture or 
slit, as important and as epoch-making as the investigations 
concerning the refraction of light. The instrument used by 
Fraunhofer was essentially a twelve-inch repeating theodo- 
lite whose verniers read to 4". In the middle of the circle 
and about it there is a plane horizontal disk six inches in 
diameter which turns on its axis, and whose center lies ex- 
actly on the axis of the theodolite. On this disk, the slit to 
be investigated was placed. The width of the opening in 
the slit was measured by a micrometer devised for the pur- 
pose which could read to o.oooi inches. Light passing 
through a narrow opening at the heliostat and falling on the 
screen with its slit is examined by the theodolite telescope. 
Fraunhofer discovered the diffraction pattern which consists 


of a bright strip in the center having symmetrically on each 
side a series of bright and dark bands, if the incident light 
is monochromatic. If the incident light is sunlight, instead 
of having bright and dark bands, there will be found a series 
of colored bands in which, however, the transitions from one 
color to another are not sharply defined. The series of 
bands gradually decrease in brightness in passing outwards 
from the central beam of light until they are finally rendered 
invisible. Instead of employing a long, narrow opening at 
the heliostat, he used a small circular opening, in front of 
which he placed a cylindrical lens and as a result obtained 
bands identical with those produced by the other method. 
In the process of development of this subject, the natural 
course after trying diffraction through a single opening, 
would be to try the action of two, three and more parallel 
openings. This was exactly the plan followed by Fraun- 
hofer. In order to study the diffraction through a great 
number of openings he stretched upon a rectangular frame 
a great many wires of the same thickness, parallel to each 
other and at the same distance apart. The light was then 
diffracted through the intervening spaces. In order to be 
sure that the wires were exactly parallel and at exactly equal 
distances apart, he made two very good micrometer-screws, 
and putting these on opposite sides of a frame, he wound 
on this frame very fine wire, being careful to stretch the 
wire at a constant tension. If now he soldered along the 
length of the screws, each wire was thus securely fastened, 
and by sawing each screw in halves, two similar wire grat- 
ings were obtained. By this method gratings were manu- 
factured consisting of wires 0.002 inches thick, and sepa- 
rated by spaces of 0.004 inches. Using a grating of 260 
turns of wire, and examining the light which first passed 
through a narrow opening at the heliostat and then fell on 
the grating placed in front of the theodolite objective, he 
found, much to his surprise, phenomena which were entirely 
different from those observed with a single opening. The 
aperture at the heliostat was seen exactly as if no grating 
intervened, but on both sides were seen a series of spectra 
as perfect as he had hitherto obtained with a good prism. 


Photographed at Mt. Wilson. The black circle at left of lower picture 
shows the size of the earth. 














The series of spectra gradually increased in length but de- 
creased in intensity, and with his apparatus he was able to 
count thirteen spectra on both sides of the middle. In order 
to vary the conditions as much as possible, gratings were 
made of different thickness of wire and with different spaces. 
Wire was wound on a screw having as many as 343 threads 
to the inch. Gratings were also made by scratching parallel 
lines on a piece of glass covered with goldfoil, through which 
spectra were observed exactly similar to those observed with 
wire gratings. Fraunhofer quickly found that the size of 
the spectra produced did not depend upon the width of the 
spaces nor upon the thickness of the wires, but upon the 
sum of these two quantities, or the distance apart of the 
centers of the wires. Consequently, the finer the screw in 
whose grooves the wires were stretched, the longer would 
be the spectra, and it became immaterial of what thickness 
the wire was or how wide the opening. The quality and 
accuracy of a grating depends on the precision attained in 
the attempt to arrange wires of the same width throughout, 
and in making the wires perfectly parallel, with their centers 
equally distant. 

Although the same dark lines were seen in the spectrum 
of sunlight when produced by the grating as were found 
when a prism was employed, one point of difference was 
revealed in the relative distances apart of the lines in the 
two spectra. In the diffraction spectrum of the grating, each 
of the different colors, the red, orange and so on through 
the blue and violet, is about equal in extent, while in the 
prismatic spectrum on the other hand the colors at the red 
end grow more and more crowded together, with the result 
that the violet reaches to a much greater extent in prismatic 
spectra than the red. Consequently, the appearance of a 
diffraction spectrum differs very much from that of a pris- 
matic spectrum, and it is well to be familiar with the two 
different types of spectra. 

In a normal spectrum produced by a grating, the disper- 
sion from the C line in the red to the D lines in the orange 
is approximately twice the distance from G to the H-line in 
the violet; whereas with a flint prism of 27 angle the ex- 


tents of the red and violet are changed in such a remarkable 
manner that the distance from C to D is only one-half that 
from G to H. 

There is also another marked difference between prismatic 
and diffraction spectra. In using a prism to form a spec- 
trum, the red rays are the least bent from their incident di- 
rection, and as a result we speak of the red as the least 
refrangible. How is it with the diffraction spectrum with the 
bright beam in the center and spectra on both sides? Is the 
violet or the red end towards the bright patch? The red is 
the least refrangible end in prismatic spectra, but not in 
diffraction spectra, for the violet is nearest to the bright 
patch with the red end bent more from the original direc- 
tion. However, since spectrum analysis was first developed 
from the use of prisms, a nomenclature has been adopted 
which, though applying essentially only to prismatic spectra, 
is used indiscriminately in respect to spectra of both kinds. 
Thus in referring to the least refrangible rays, the red end 
of the spectrum is always meant and never the violet, al- 
though in the diffraction spectrum the violet is the least 
refrangible end. 

After measuring the angles by means of the theodolite, 
Fraunhofer was able to formulate two laws: first, the size 
of the spectra, and their distances from the center (or the 
dispersion) vary inversely as the distance between the cen- 
ters of the lines in the grating; and second, the dispersions 
in the different spectra for any ray form an arithmetical 

This second law states in other words that the angle of 
deflection of the same colored beams in the series of spectra 
formed by the grating are in the ratio of the numbers i, 2, 
3, etc. The experiments from which these results were de- 
duced gave, however, such small angles that the sine, the 
tangent and the arc do not sensibly differ. If the angles 
were larger, that is, if the gratings had greater dispersion, 
it might be possible to determine whether it was the arcs 
themselves that form an arithmetical series, or some func- 
tion of these angles. Having this in view, Fraunhofer made 
further experiments on gratings to find whether it would not 


be possible to get gratings of greater dispersion. As it was 
almost impossible to evolve a screw, for the manufacture of 
his wire gratings, with a smaller pitch than the one he had 
already employed of 343 threads to the inch, he constructed 
a machine for scratching parallel lines upon a piece of plane 
glass covered with gold-foil, and he succeeded in scratching 
them so closely together that he was able to rule a grating 
with about 900 lines per inch. If more lines than this were 
scratched, practically no gold- foil remained. With this grat- 
ing and with others, he measured the deflection for different 
rays in different orders of spectra and he found that the 
sines of the angles of deflection increased uniformly in the 

different orders of spectra, or in other words Sin = -, 


where is the angle of deflection, n is the order of spectrum, 
X is the wave-length and co the grating space. 

Using this equation, Fraunhofer was able to astonish the 
world by telling them that he had been able to measure the 
infinitesimal length of a light-wave, for the D line of 
0.00005888 cm., which is very close to modern determi- 

To sum up the work of Fraunhofer: he was a telescope 
maker, and in order to improve his lenses entered into an 
investigation of prisms in order to study the action of light 
in passing through them. Using a telescope in connection 
with his apparatus, he found the spectrum of sunlight filled 
with lines, now called " Fraunhofer lines." Investigating 
the spectra of flames he found the bright D line doubled and 
in the same position exactly as the dark D in the solar spec- 
trum. He examined the spectra of planets and stars. Tak- 
ing up diffraction, he studied the action of light passing 
through a single opening, and formulated the laws govern- 
ing it. He then studied the action of light through parallel 
openings side by side. Next he attempted to rule lines in 
a layer of grease spread over a glass plate so thinly that the 
film could scarcely be recognized by the eye alone. In this 
grease parallel lines were scratched which were only half as 
far apart as the lines ruled in gold-foil. After many experi- 
ments it was found impossible to rule lines in any layer of 


grease or varnish much finer than the gold-foil grating. A 
diamond point is the only method yet known in modern en- 
gineering that will provide sharp and sufficiently clean cut 
lines to permit the construction of finer gratings. 

Let us stop a moment and think what it means to rule a 
grating with a diamond and try to visualize the total lengths 
of the lines the diamond point must trace, practically with- 
out variation. The finest gratings ruled by Rowland have 
20,000 lines per inch, and if the grating possesses a ruled sur- 
face of 3 x 6 inches, the diamond evidently rules 3x6 
x 20,000 or 360,000 inches, a distance of nearly six 
miles. If in ruling this distance the diamond point 
wears appreciably or breaks down, the previous rulings 
are useless. Altogether it takes from five to six days 
continuous working to rule such a grating. What a task 
Fraunhofer must have had in ruling his gratings! In- 
deed it must have consumed an enormous amount of 
time and patience, with numerous vexations caused by 
imperfect rulings and the fracture of the diamond points. 
It was only after many trials that he succeeded finally in 
getting a grating with about 7500 lines to the inch. In 
order to obtain good definition with his little telescope it 
was necessary for the slit to be at some considerable dis- 
tance (as much as 642 feet in one set of experiments) in 
front of the prism. A great advance in instrumental equip- 
ment was later made by Simms, of the celebrated optical firm 
of Troughton and Simms, who rendered, by the introduction 
of a lens between slit and prism, the incident rays parallel 
in falling on the prism. Thus was introduced the collimat- 
ing lens which increased the compactness of the spectroscope 
and rendered it almost as we now use it. 

In England, Brewster and Herschel undertook a great va- 
riety of experiments by means of which they examined the 
absorbing effects of various colored substances. A piece of 
red glass permits only the red part of the spectrum to pass 
through it, the balance of the spectrum being absorbed by 
the glass, while blue glass allows only the blue end of the 
spectrum to pass unobstructed. It consequently became 
evident that the absorbing effect of the terrestrial atmos- 


phere should be carefully investigated before any certain 
information could be obtained regarding the cause of the 
dark lines in the solar spectrum. The water vapor in our 
atmosphere exerts a powerful absorbing action on the light 
of the sun as it passes through the air on its way to reach 
the slit of the spectroscope. The absorption commences at 
the blue end of the spectrum and becomes gradually greater 
and greater as the sun sinks towards the horizon. For very 
dense layers of atmosphere, when the sun is near the horizon 
either in rising or setting, the absorbent action is s5 great 
that little of the solar spectrum penetrates except the yellow 
and red parts, the violet end being entirely absorbed. For 
this reason the sun is always red when rising or setting, and 
the color becomes a deeper red when dust or haze in the 
atmosphere causes a greater absorbing effect. At the time 
of a total eclipse of the moon, the light from the sun passes 
through what might be termed a double layer of terrestrial 
atmosphere with the consequence that there is an increased 
absorption, the moon thus shining with a dull copper colored 
hue, and this in spite of the fact that the moon is immersed 
in the shadow cast by the earth! 

By 1833, Brewster announced that he had examined the 
lines of the solar spectrum with various optical combina- 
tions, and had made a map of the solar spectrum on a scale 
four times that of Fraunhofer. Indeed for some portions 
of the spectrum the scale was twelve times Fraunhofer's. 
By observing the absorbing effect of various substances he 
soon found that some of these substances exerted a general 
darkening action on the spectrum, whereas other materials 
produced an absorption in a limited portion of the spectrum. 
At times the effect was so limited in action that bands or 
even lines, sharp and distinct, were added to the solar spec- 
trum. These phenomena were so clearly defined that the 
conviction was borne in upon Brewster that he had obtained 
" the discovery of a general principle of chemical analysis, 
in which simple or compound bodies might be characterized 
by their action on definite parts of the spectrum.' 7 

In the course of his investigations, Fraunhofer varied the 
conditions under which he observed, using different kinds 


of slits to see if the dark lines in the solar spectrum were 
caused by some action at the slit itself; but when he found 
that the form of the slit had nothing to do with the presence 
of the lines, he concluded that the lines must be truly solar 
in their origin. Brewster endorsed this view and inferred 
that if a tube of nitrous oxide gas gave lines identical in 
character with the solar lines, these lines must then be 
caused by absorption at the surface of the sun. When in 
addition he found that many of the lines of nitrous acid gas 
appeared to be identical in position with some of the Fraun- 
hofer lines, the verification seemed to be complete. Brewster 
made another important discovery which is described in 
his own words as follows : " When the sun descends towards 
the horizon and shines through a rapidly increasing depth of 
air, certain lines which before were little, if at all, visible, 
become black and well defined, and dark lines appear even 
in what were formerly the most luminous parts of the spec- 
trum." As these lines appear both at sunrise and sunset, 
Brewster announced the discovery that these bands and lines 
were caused by the absorbent effect of the earth's atmos- 
phere. Since the majority of the lines in the sun, however, 
appeared without change, he concluded that, " the apparent 
body of the sun is not a flame in the ordinary sense of the 
word, but a solid body or coating raised by intense heat to a 
state of brilliant incandescence." 

To quote from Lockyer, Chemistry of the Sun, page 41, 
" It will be seen, then, that the study of the sun was now 
(1833) in full swing. We had at length, after waiting some 
centuries, a method of observing a spectrum; we had, 
further, the fact that there were dark lines in the solar 
spectrum; that colored flames gave us bright lines; that 
certain substances stopped some of the light which passed 
through them, thus producing dark lines. Hence that the 
solar lines might be produced in the same way." 



THE solar eclipse of 1836 played a very important 
role in the history of the development of the new 
science of astrophysics. The eclipse was visible at 
Edinburgh. It was not a total eclipse, however, but merely 
an annular one. At such an eclipse the angular diameter of 
the moon is smaller than that of the sun, with the result that 
at the middle of the eclipse there is a ring or annulus of sun- 
light visible around the edge of the dark moon. This eclipse 
was observed by Forbes. Referring to Brewster's discovery 
of atmospheric lines in the solar spectrum, Forbes clearly 
pointed out that the " telluric " lines were comparatively 
few in number and relatively unimportant compared with 
the very great number of solar lines. Moreover, he empha- 
sized the fact that the Fraunhofer lines could not have been 
caused by the absorbing effect in our terrestrial atmosphere, 
for if this were true the spectra of the stars should be iden- 
tical with that of the sun. As the spectra of the stars dif- 
fered among themselves and were not always identical with 
the spectrum of the sun, it was manifest that it was not pos- 
sible to imagine a terrestrial origin for all of the solar lines. 
Consequently, if the absorbing action of the earth's atmos- 
phere could not be invoked to give an adequate explanation, 
there remained no other cause than that the origin of the 
lines must take place within the sun's own atmosphere. If, 
therefore, this was the true explanation, then an annular 
eclipse of the sun should furnish a crucial test. The ter- 
restrial lines of the solar spectrum become more and more 
intensified as the sun's light reaches us through greater and 
greater layers of our earth's atmosphere, and in like fashion 
it appeared evident to Forbes that a similar effect must be 
visible in the lines of truly solar origin. And since the light 



from the limb of the sun must pass through a much greater 
thickness of solar atmosphere than that from the sun's cen- 
ter, the lines from the sun's edge, observable at the time of 
the annular eclipse, should be much intensified in comparison 
with the Fraunhofer lines ordinarily visible from the sun's 
center. When the eclipse took place no change whatever 
was observed by Forbes in the number, position or intensity 
of the lines, and the obvious conclusion was drawn by him 
that, " This result proves conclusively that the sun's atmos- 
phere has nothing to do with the production of this singular 

To us living in the twentieth century, it seems passing 
strange that Forbes did not try the experiment of comparing 
the limb and the center of the sun by forming an image of 
the sun on the slit of his spectroscope by means of a project- 
ing lens. This plan was undoubtedly in his mind as the 
following shows: "Had the weather proved unfavorable 
for viewing the eclipse, I intended to have tried the experi- 
ment by forming an image of the sun by using a lens of long 
focus, stopping alternately by means of a screen the interior 
and central moiety of his rays, and restoring the remainder 
to parallelism by means of a second lens, then suffering these 
to fall on the slit as before. The result of my experiment 
during the eclipse seemed, however, so decisive as to no 
marked change being produced at the sun's edges that I 
have thought it unnecessary to repeat it." 

The scientific world in the first half of the nineteenth 
century knew little of laboratory methods and it was not 
at that time a habit of mind to test any theoretical conclu- 
sions by means of experimentation. In fact it was not until 
the year 1866 that the idea of Forbes was carried into execu- 
tion and a projecting lens utilized for the examination of 
local phenomena such as the spectra of the limb or of sun 

In 1845, there was performed in England by W. H. Miller, 
the very experiment that later in Kirchhoff s hands was the 
crucial proof of the cause of the Fraunhofer lines. This 
experiment consisted in passing sunlight through various 
vapors heated to incandescence in a flame and noting the 


changes in the solar spectrum. He observed that in passing 
sunlight through glowing sodium vapor, the D lines were 
intensified. It is surprising that he appears to have sought 
no explanation of this amazing fact, and the surprise is all 
the greater since Miller's experiments were carried out with 
the express purpose of testing the theory that the Fraunhof er 
lines were actually produced by absorption in the sun's 

This line of investigation was continued in 1849 by 
Foucault in Paris who was able to use a new method of 
heating salts and metals to the glowing point by the use of 
the electric arc. He focused an image of the sun on the 
arc itself and this procedure allowed him to observe at the 
same time the spectra of the arc and of sunlight. Inciden- 
tally he was surprised to discover the extreme transparency 
of the arc which caused only a faint shadow in the sunlight. 
This experiment manifested to Foucault that when the two 
spectra were exactly superimposed, the D line of sunlight 
was made considerably darker, proving that the arc ab- 
sorbed the D rays; but that when, on the contrary, the two 
spectra jutted out one beyond the other, the D line appeared 
darker than usual in sunlight, yet stood out bright in the 
electric spectrum, thus demonstrating the perfect coinci- 
dence in position ot the dark and bright rays. " Thus the 
arc presents us with a medium which emits the rays D on 
its own account, and which at the same time absorbs them 
when they come from another quarter." It was difficult to 
explain how it was possible for a glowing flame to furnish 
at the same time both bright and dark lines, and the riddle 
of the solar lines seemed a hard one to solve. As early as 
1850, Stokes seemed to have clearly grasped the solution 
of the problem, and he inserted a discussion of these matters 
in his university lectures at Cambridge. 1 Moreover, he 
seems to have been the first to localize the cause of the 
yellow D lines since he observed that the bright line was 
absent from a candle flame when the wick was snuffed clean, 
and from an alcohol flame when the spirit was burned in a 
watch-glass. Although he so clearly saw that the D lines, 

1 See Lockyer, Chemistry of the Sun, p. 51. 


whether bright or dark, were caused by sodium vapor, and 
therefore correctly concluded that the absorbing effect of 
the sodium was in the neighborhood of the sun, he made no 
further experiments to test his deductions; and the honor 
for the great discovery waited another nine years. Thus 
again was demonstrated which has frequently happened 
in the history of science that though many investigators 
have converged towards the same goal of discovery, yet the 
prize has awaited the fortunate one who should make the 
critical experiment, which in itself at times has been one of 
little difficulty. Thus in 1855 Angstrom, and in 1859 Bal- 
four Stewart by their experiments came very close to the 
true solution of the problem. 

In 1859, Kirchhoff showed ] for the first time that in order 
that sodium should be in a condition to absorb from Uie 
light of other sources it must itself be at a cooler tempera- 
ture. " Fraunhofer has remarked that in the spectrum of a 
candle flame two bright lines appear which coincide with the 
two dark lines D of the solar spectrum. These bright lines 
can be easily intensified in a flame into which some common 
salt is put. I formed a solar spectrum by projection and I 
allowed the solar rays thus formed to pass through a strong 
salt flame before falling on the slit. If the sunlight were 
sufficiently subdued, then in place of the two dark lines D 
two bright lines appeared ; if the intensity increased beyond 
a certain amount then the two dark D lines showed in much 
greater intensity than without the presence of the salt flame. 
The spectrum of the Drummond light contains as a rule the 
two bright sodium lines if the illuminating spot of the cal- 
cium cylinder has not long since passed the glowing point; 
if the cylinder remains undisturbed then these lines become 
weaker and finally completely vanish. If they have dis- 
appeared or are faintly visible, an alcohol flame into which 
cooking salt has been placed and which is brought between 
the calcium cylinder and the slit, causes two dark lines of 
exceptional blackness and sharpness, which in that respect 
agree with the lines D of the solar spectrum, to show them- 
selves in their place. In this manner the D lines of the solar 

1 For complete details see Kayser, Handbuch der Spectroscopie, Vol. T, p. 81, 


spectrum are artificially produced in a spectrum in which 
they are naturally not present. I conclude from these ob- 
servations that colored flames, in the spectra of which bright 
sharp lines are found, so weaken rays of the color of these 
lines when such rays pass through the flames, that in place 
of the bright lines, dark ones appear just as soon as there is 
brought behind the flame a source of light of sufficient in- 
tensity in the spectrum of which these lines are otherwise 
lacking. I conclude further that the dark lines of the solar 
spectrum, which do not find their origin in the earth's atmos- 
phere, are caused in the glowing solar atmosphere by the 
action of those substances which in the spectrum of a flame 
produced bright lines at the same place. We thus assume 
that the bright lines coinciding with D in the spectrum of 
a flame always arise from sodium contained in it; the dark 
D lines in the solar spectrum therefore allows us to conclude 
that sodium is found in the atmosphere of the sun. ... In 
order that the D lines should come out dark in the spectrum 
of the Drummond light it is necessary to use a salt flame 
of lower temperature." 

In these experiments, the salt flame was kept constant 
the intensity of the sunlight being weakened or strength- 
ened at leisure. 1$ other words, this may be expressed by 
saying that the Fraunhofer lines in the spectrum of the sun 
are dark only in contrast with the more brilliant background 
of the sun itself. If this dazzling surface could be removed, 
and the comparatively dark lines of the solar spectrum could 
then be viewed against a background still darker, then by 
contrast the spectrum lines would appear as bright lines on 
a dark background where formerly they had existed as dark 
lines on a bright background. This change in the sun's 
spectrum, as we shall see later, takes place at the time of 
a total eclipse of the sun. The student of spectroscopy will 
save himself needless worry if he will remember that bright 
and dark are always to be considered as relative terms only. 

A spectrum of bright lines on a dark background is said 
to be a bright-line, or an emission spectrum. On the other 
hand, a spectrum of dark lines on a bright background is 
called a dark-line, or an absorption spectrum. If a chemi- 


cal element is heated to the point of vaporization, its spec- 
trum consists of a bright-line spectrum. Sodium gives a 
very simple spectrum, consisting mainly of two very strong 
lines in the yellow part of the spectrum, the well-known D 
lines. If pure metallic sodium alone is used, the spectrum 
consists of these D lines. If the sodium is in chemical com- 
bination with chlorine, and the sodium chloride, or common 
salt, is heated to incandescence, the same D-lines due to 
sodium are shown in the bright-line spectrum. Or if any 
other compound of sodium is heated, the same D-lines re- 
sult. The manner of the heating is of no consequence; a 
pinch of common cooking salt may be placed on the wick of 
an alcohol flame, a paper soaked in a saline solution may be 
placed about the burner of a Bunsen lamp, or a grain of 
salt may be put on the carbon of the electric arc or on one 
of the poles of an electric spark the lines of sodium will 
always appear, the color of the lines will be exactly the 
same, and the wave-lengths of the lines will be unaltered 
no matter what the chemical compound in which the sodium 
is found or the manner of heating that salt to incandescence. 
Since the total light of the sodium consists mainly of two 
lines in the yellow, then when sodium is burning,, it will 
give off yellow light only and the color of the flame will 
appear yellow to the eye. 

If another element like lithium is examined, whether pure 
lithium or lithium in compound with some other element or 
elements, it will give its own peculiar spectrum of bright 
lines, and these bright lines will be found not to coincide 
with the D lines due to sodium. Since lithium burns with a 
red light, the prominent lines in its spectrum will be found 
at the red end of the spectrum. Some metals like sodium 
show a very simple spectrum, with very few lines; other 
metals show more lines, the greatest number of lines appear- 
ing for any one element being due to the presence of iron. 
It makes no difference how the iron is heated to incandes- 
cence, it makes no difference whether the iron is a piece of 
scrap or of polished steel, the spectrum will consist of 
thousands of bright lines in all the colors of the spectrum. 
Each of these many thousands of lines has its own particular 


wave-length. It is the business of the spectroscopist to find 
the value of the wave-length of each and every line. Some 
of the lines are faint, some strong, some are narrow, some 
broader, some are very sharp, others more fuzzy in appear- 
ance, but no matter what the quality of the iron or how 
vaporized, the spectrum is the same with lines of practically 
identical wave-lengths. Although there exist a great many 
chemical elements, the spectrum of each of which consists 
of many lines, while others of the elements have even thou- 
sands of lines in their spectra, it may almost be said that 
no line in any one spectrum coincides precisely with a line 
in any other spectrum. If, therefore, it is possible to de- 
termine the exact wave-length of a line in a spectrum, 
though the chemical origin of this line may be unknown, 
we shall have a ready means of identifying the elemental 
source of this line. 

The principles upon which spectrum analysis depends are 
found in Young's General Astronomy, page 213, as follows: 

1. A continuous spectrum is given by every incandescent 
body, the molecules of which so interfere with each other 
as to prevent their free, independent, luminous vibration; 
that is, by bodies which are either solid or liquid, or if gase- 
ous, are under high pressure. 

2. The spectrum of a gaseous element, under low pressure, 
is discontinuous, or in other words made up of bright lines, 
these lines being characteristic, that is, the same substance 
under similar conditions always gives the same set of lines, 
and generally does so even under widely different conditions. 

3. A gaseous substance absorbs from white light passing 
through it precisely those rays of which its own spectrum 
consists. The spectrum of white light which has been trans- 
mitted through it then exhibits a " reversed " spectrum of 
the gas; that is, one which shows dark lines instead of the 
characteristic bright lines. 

The third law, the great discovery of Kirchhoff, may be 
stated in other words as follows: The relation between the 
emissive power for each wave-length and the absorptive 
power for the same wave-length at the same temperature is 
identical for all bodies, and is in fact equal to the emissive 


power of an absolutely black body at the same wave-length 
and temperature. We are quite familiar with a similar effect 
in the realm of sound. If a voice singing or speaking sounds 
a note of a certain pitch in a room where there is a piano, 
one string of the piano will vibrate in unison with the voice, 
the particular piano string taking its motion from the oscil- 
lations of the air. Similarly, a tuning fork in vibration will 
set in motion another tuning fork nearby which is tuned to 
the same pitch. 

The work of Kirchhoff, therefore, in connecting the emis- 
sion of light with absorption gives the means of determining 
the chemical composition of the sun. According to our 
present ideas, the photosphere of the sun, the portion of the 
sun we see, consists of gases under such very high pressure 
that the molecules cannot vibrate independently; and in 
consequence the spectrum of the photosphere must be con- 
tinuous, a ribbon of light without breaks from red to violet. 
The photosphere is surrounded by a cooler layer of gases 
under low pressure, the so-called " reversing layer." If the 
spectrum of these gases could be examined entirely sepa- 
rated from the bright photospheric background, they would 
exhibit the gaseous, or bright-line spectrum. Under ordi- 
nary conditions, the light from the photosphere shines 
through the cooler gases of the reversing layer, and certain 
wave-lengths are there absorbed by the gases of the revers- 
ing layer, so that the spectrum of the sun comes to us as a 
reversed spectrum, of dark lines on a bright background; 
these dark lines, however, as stated above, are dark only in 
contrast with the much brighter photospheric background. 
To determine the constitution of the sun, it becomes there- 
fore necessary to compare the bright line spectra of the va- 
rious elements with the dark line spectrum of the sun. This 
comparison may be made by two different methods, either 
by viewing or photographing with a suitable instrument the 
spectrum of the sun and the comparison spectrum side by 
side, or by an exact determination of wave-lengths in the 
solar and in the comparison spectrum. 

But the spectrum of the sun consists of many thousands 
of lines. It is evidently quite possible, and even highly 


probable, that there should be very close agreement be- 
tween some of the many lines in the sun and an equal num- 
ber of lines in the spectrum of the element under investiga- 
tion. These concurrences might be the result of pure acci- 
dent. Kirchhoff investigated this possibility. A particular 
line in a comparison spectrum may exactly match in position 
a line in the solar spectrum. If the agreement is due en- 
tirely to chance, then by the laws of probability, it is equally 
probable that the line in the spectrum under consideration 
may or may not match a line in the sun's spectrum, or speak- 
ing mathematically the chance of an exact match taking 
place fortuitously is one out of two. If two lines agree in 
each spectrum, then the possibility of this happening by 
chance is but one out of four. Kirchhoff found sixty lines 
of iron to agree with sixty lines in the sun. The chance that 
this coincidence of all sixty lines is purely accidental is ex- 
pressed by the number ^ raised to the sixtieth power. At 
the present time over two thousand lines due to iron have 
been identified in the solar spectrum. If the coincidence of 
this large number of lines were the result of pure accident 
it would represent a chance of one in 2 raised to the 2Oooth 
power. This number is about equivalent to 100 followed by 
no less than 600 ciphers! (If one has nothing better to do, 
one might take a large piece of paper and put down the num- 
ber one and follow it by six hundred and two zeros. Then 
one could divide it off into millions, billions, trillions, etc., 
and invent a name for this huge number ! ) The chance that 
the lines of iron and the Fraunhofer lines in the sun should 
agree in position entirely by accident is therefore infinitesi- 
mally small. But when in addition, we compare the appear- 
ance of the lines in the two spectra, and find that a strong 
line in the spectrum of the sun is matched by a strong line 
in the spectrum of iron, and a weak Fraunhofer line is 
matched by a weak iron line, then we see the utter impossi- 
bility of the coincidences being the result of mere chance. 
What is true of iron is equally true of the other elements 
investigated. It accordingly seems perfectly certain that 
we are able to ascertain the chemical constitution of the sun 
by means of the spectroscope even though we are looking 


at the sun across a space of ninety-three millions of 

Astrophysics thus being placed on a very firm foundation, 
the infant science was immediately recognized throughout 
the scientific world to be of the very greatest importance to 
physicists, chemists and astronomers. But in spite of the 
almost universal recognition, there were a few doubting 
Thomases. In Chambers's very excellent Descriptive As- 
tronomy, edition of 1867, page 27, is found the following: 
" Spectrum analysis has taken a start within the last two or 
three years, chiefly owing to the assertions made that it 
enables us to ascertain something about the physical condi- 
tion of the sun. The subject is too purely a physical one, 
and also in too infantine a state to require notices in these 
pages at present, though the time may come/' 

It is remarkable to read the early history of spectroscopy 
and learn of the great opposition of certain scientists to 
the acceptance of Kirchhoff's proof; and it was but 
natural that there should be many claims to priority. 
By 1867, however, Pritchard of Oxford summarized the 
general feeling of the scientific world in the following words : 
" It may safely be asserted of Foucault in 1849, of Stokes in 
1850, of Angstrom in 1855, and of Balfour Stewart in 1859, 
that each of them was in possession of an enunciated truth, 
which, had they traced to their natural and inevitable con- 
sequences, must have led to that grand generalization which 
will immortalize the name of Kirchhoff, and which forms 
one of the happiest and most remarkable discoveries of 
modern times/ 7 

With the trail so surely blazed, the path pointed out the 
direction of the researches to be taken by Kirchhoff's suc- 
cessors. The quest had a two-fold interest, for not only 
did the new method serve as an infallible chemical test of 
terrestrial substances, but it gave also a ready means of 
determining the constitution of the sun and also of the more 
distant suns, the stars. Physicists, astronomers and chem- 
ists vied with each other in pushing forward the researches 
as rapidly and thoroughly as possible, and opticians and 
instrument makers came to the assistance of the scientists 


by furnishing improved forms of apparatus. It was mani- 
festly necessary to investigate the solar spectrum and the 
spectra of the various chemical elements. These investiga- 
tions had necessarily to be carried out with the spectra, 
produced on as large a scale as possible, and the measure- 
ments of the positions of the lines required the very highest 
degree of precision attainable. To denote the position of a 
line, some more accurate method was necessary than that of 
placing it in the red or blue of the spectrum. Fraunhofer 
and also Kirchhoff used a rather arbitrary scale. Newton 
proved that a difference in color meant a difference in re- 
frangibility. Fraunhofer went a step further and demon- 
strated that a difference in color meant a difference in the 
length of the wave causing the light. Since the time of 
Thomas Young, 1802, it has been known that light is a 
wave phenomenon, the waves, somewhat similar to water 
waves, moving transversely to the direction of motion. 
The length of the wave from crest to crest, or from trough 
to trough, is known as the wave-length, and this length 
might be measured in fractions of an inch, or foot, or meter. 
All scientists, whether they live in America or in Germany, 
or whether they speak English, Japanese or Russian, now 
use the meter as the unit for measuring the wave-length 
of light. A meter divided into ten thousand million parts, 
or 10 10 parts, is called a " tenth-meter." This very small 
distance is known as the " Angstrom Unit," or more simply 
as the " Angstrom," and it is the unit for measuring wave- 
lengths. The position of a line is known by its wave-length, 
and the more precise the investigation the more accurately 
do we need to know this quantity. The K-line in the solar 
spectrum has a wave-length according to Rowland of 3933.- 
826. This is printed either as X 3933.826, or 3933.826 A. 
(We shall adopt the latter notation.) 

Knowing the wave-length of light of a certain color, it 
is a very simple matter to calculate the number of waves 
that enter into the eye in a single second of time. All that 
is required is to know the velocity at which light travels. 
It is now known that light of all colors, whether red, blue or 
violet, travels at the same rate of speed, viz., the almost 


incredible velocity of 186,300 miles, or in round numbers 
300,000 kilometers per second. To simplify the calculation, 
suppose the light is violet, of wave-length 4000 A. The 
length of these waves from crest to crest is 4000x10"', 
which is 7 meters (i.e. 4 divided by ten million), 
300,000 kilometers per second is 300,000,000 meters 01 8 meters per second. If therefore we divide the dis- 
tance that light travels in one second by the wave-length , 
we will find the number of waves. For light of 4000 A, 
seven hundred and fifty millions of millions (750,000,000,- 
000,000) of waves enter into the eye in one second of time. 
If the light under consideration is red instead of violet, in- 
asmuch as the wave-length of the red is longer than that 
of the violet, fewer red waves will consequently enter into 
the eye in a given time. These tiny waves impinge on the 
retina of the eye, creating motions which when telegraphed 
to the brain cause the sensation of violet or red light. The 
mechanism by means of which the minute motions produced 
in the eye by light waves cause the sensation of light has 
never been completely discovered. Professor John Joly in 
Philosophical Magazine, 42, 289, 1921, gives a very plaus- 
ible explanation based on the quantum theory. He as- 
sumes that the origin of vision and color perception is to 
be sought in the liberation of electrons under light stimulus 
within a photoelectric substance or substances existing in 
the retina. In the case of the rods, rhodopsin, being such 
a photosensitive substance, acts as the basis of vision, and 
it is assumed that the same substance in the cones is re- 
sponsible for the color vision. The sensitivity of the eye 
to faint light is extraordinary. Henri and des Baucels have 
found that the retina is sensitive to a minute amount of 
light 'energy which, when expressed in physical units, 
amounts to 5xio~ 12 erg. The quantum for green light is 
4xio~ 12 erg, and hence it is assumed that one quantum, by 
the liberation of a single electron, is sufficient to cause the 
sensation of light. The action taking place in the eye 
seems to be quite analogous to that occurring in the light- 
sensitive film of the photographic plate, the latent image 
being caused by the movement of electrons. 


The new method of research in the hands of Kirchhoff 
soon resulted (1861) in the discovery of two new chemical 
elements, caesium and rubidium. The spectroscope manu- 
factured by Steinheil consisted of four prisms, three of 45 
and one of 60. The collimator and telescope had aper- 
tures of one and a half inches, with a focal length of eight 
een inches. One half of the slit was covered by a totally 
reflecting prism by the aid of which two spectra could be 
examined side by side and direct comparisons made. Un- 
fortunately, it was necessary to set to minimum deviation 
by hand, a very slow process, and as a consequence Kirch- 
hoff shifted the prisms only occasionally, a procedure which 
greatly impaired his measures. He investigated the spectra 
of a large number of elements, and also measured the posi- 
tions of the lines in the solar spectrum from A to G, though 
at first he only published the region from D to F, since his 
eyes could not stand the strain of such continuous measure- 
ment and failed him. His measures being referred to an 
arbitrary scale, it was necessary to reduce them to wave- 
lengths, and this was done later by Airy, Gibbs, Watts and 

It has been said with great verity that these were splen- 
did days for the laboratory scientist for the reason that 
each and every observation, no matter how trivial, was 
almost certain to prove to be a new discovery. Without 
attempting to trace the details in the further development 
of the new science, we shall try to give only the more im- 
portant names in the honor roll of fame: Pliicker, Hittorf, 
Crookes, Miller, Huggins, Rutherfurd, Angstrom, Secchi, 
Janssen, Lockyer, Young, Vogel, Cornu, Liveing and De- 
war. In the twenty years following Kirchhoff, no less than 
ten new elements were found by the aid of the spectroscope. 
Naturally many mistakes were made and wrong conclusions 
drawn. The chief cause of the mistakes was the presence 
of many impurities in the elements investigated and hasty 
identification of lines through insufficient accuracy in wave- 
length determinations. 

A new epoch in the history of spectrum analysis was in- 
augurated in 1882 by the work of Henry A. Rowland, whose 


gratings/ plane and concave, permitted a hundred-fold in- 
crease in accuracy in the determination of wave-lengths. 
Rowland's success came through the construction of a long 
screw, almost free from errors, mounted in a dividing engine 
in such a manner that it was practically possible to elimi- 
nate the few remaining errors of the screw. The precision 
that must be attained in the manufacture of gratings of 
the very first quality may be stated as one requiring that 
the average line of the grating shall be correctly placed to 
about one one-thousandth part of the grating space. This, 
for the finest Rowland gratings of 20,000 lines per inch, 
means that each line ruled on the grating must not on the 
average differ from its true position by so much as the 
minute quantity of one twenty-millionth part of an inch! 

The highest degree of success was attained by ruling with 
a diamond point on speculum metal, the incident light thus 
being reflected from the grating surface. Gratings were 
ruled both on plane and on spherically concave surfaces. 
The largest Rowland gratings were six inches in diameter, 
and ordinarily the greatest radius of curvature for the con- 
cave gratings was twenty-one and a half feet. The concave 
gratings reduced the spectroscope to the greatest simplicity 
of slit, grating and photographic plate. No lenses of any 
kind were necessary to bring the light to a focus, and hence 
all aberrations introduced by the lenses and all absorption 
of light by the glass were eliminated, with a consequent 
great increase in the extent of the ultra-violet region. The 
concave grating is generally used in the laboratory with the 
" Rowland mounting " possessing two tracks for carrying 
grating and photographic plate perpendicular to each other, 
the slit being placed accurately at the intersection of the two 
tracks. If grating and photographic plate are each perpen- 
dicular to the arm joining the two, there then results a " nor- 
mal " spectrum, or one in which the distances between the 
lines are directly proportional to the wave-length. 

Compared with prisms, concave gratings have the follow- 
ing advantages: (i), An enormous increase in dispersion, 
definition and resolving power. To equal Rowland's grat- 

1 See Kayser, Handbuch der Spectroscopie, Vol. i, 121 and 397. See also 
article Screw, Encyclopaedia Britannica. 


ings in these respects, in the neighborhood of the D lines, 
it would be necessary to have prisms of the very first quality 
added to prisms with a total prism base of fifty inches of 
glass. (2), The spectrum produced by the grating is 
" normal " and not " prismatic/' thus permitting wave- 
lengths to be determined with much greater facility 
and much greater accuracy. (3), A much greater extent 
of the ultra-violet is secured. (4), The astigmatism 
of the grating increases the length of the lines, and this 
principle, combined with the overlapping of images from 
the spectra of different orders, not only permits a great 
increase in accuracy, but also a more ready determina- 
tion of absolute wave-lengths. The grating, however, 
has some disadvantages, chief among which is that the in- 
cident light is divided between the central beam and many 
different orders of spectra, with the consequent result that 
there is a great weakening of light in any one spectrum. 
In the investigation of objects giving little light, like the 
stars, prisms are almost universally used. In most labor- 
atory researches and in work on the sun, gratings are gen- 
erally used. Although Rowland's method of grinding the 
screw was not a secret, it is only in comparatively recent 
times that the excellence of the Rowland gratings has been 
equalled by others, by Michelson of the University of 
Chicago, by the Mount Wilson Observatory, and by Lyle 
and Merfield of Australia. 

Coincident with Rowland's manufacture of the grating 
came the discovery of the modern photographic dry plate 
with its great increase in sensitiveness. The two most im- 
portant pieces of work in this new epoch of discovery in 
astrophysics have been Rowland's great map of the solar 
spectrum, and the publication in the Astrophysical Journal 
of the wave-lengths of the lines in the solar spectrum with 
the tracing of as many lines as possible to their chemical 
origins. The more prominent names connected with inves- 
tigations in the laboratory and on the sun itself are: Row- 
land, Jewell, Kayser, Runge, Paschen, Rydberg, Eder and 
Valenta, Exner and Haschek, Langley, Abbot, Schumann, 
Lyman, Humphreys, Zeeman, Hale, Deslandres, St. John, 
King, Meggers, Russell, Saunders and Fowler. 



AT THE time of the Greeks how simple it was to 
explain all of the known facts about the sun! No 
supposition was necessary other than that the sun 
was a ball of fire. It was not at all known what fire was 
beyond the fact that heat was manifested, but this de- 
ficiency in knowledge seemed of little importance. As a 
further illustration of the elemental beliefs of primitive 
peoples there might be given the following legend regard- 
ing the origin and motion of the sun which is found among 
the Yuki tribe of American Indians. 1 

" In the beginning there was no land; all was water. 
Darkness prevailed everywhere. Over this chaos of dark 
water hovered On-coye-to who appeared in the form 
of a beautiful white feather, hence the love of the Yukis 
for feathers. In time the spirit became weary of his inces- 
sant flight through the murky space and lighted down upon 
the face of the water. Where he came in contact there was 
a whirlpool that spun his body round and round. So rapid 
became the motion that a heavy foam gathered about him. 
This became more dense and expanded in width and length. 
It gathered up the passing bubbles until it was a huge float- 
ing island. On the bosom of this rested the snowy form 
of On-coye-to. As he lay upon this island for an almost 
endless flight through the dark space, the idea of a perma- 
nent resting place came into his mind. So he made the land 
and divided it from the water. From the form of a feather 
he assumed that of a man, and rested upon the land. Still 
there was no light, and his spirit was troubled. On-coye- 
to saw afar off in the firmament a star, ' po-ko-lil-ey/ and 
resolved to visit it and learn how it emitted its sparkling 

1 Smithsonian National Mus. Report, 326, 1902. 


light. After a long journey, he arrived there and found a 
large and beautifully lighted world, inhabited by a numer- 
ous, hospitable people. Still, he saw not whence came the 
light. He was allowed free access to all habitations save 
one, the ' sweat house. 7 This was guarded night and day, 
and was accessible only to sick persons. Finally a great 
hunt was planned, and as time drew near all was prepared 
for the occasion. But On-coye-to feigned sickness that 
he might investigate the sweat house. When the morning 
arrived for the hunt he was too ill to acompany the hunters. 
A council was held to determine whether this stranger 
should be admitted to the sweat house, which is even now 
a sacred place with the Yuki tribe, and it was decided to 
give him the benefit of this house of medicine. A few old 
men were left to administer to his wants and to see that 
all went well. As he entered the sweat house he was almost 
blinded by the light that flashed upon him, but as he became 
accustomed to it, he looked around him and discovered its 
origin. Hanging high over his head in several baskets were 
as many beautiful suns. Having found the fountain of 
light he waited patiently until the old men were all asleep, 
then climbing cautiously to what seemed the brightest of 
the suns, he seized it, slipped from the sweat house and 
made his way rapidly towards his own world. He was 
hotly pursued by the indignant warriors, but he arrived 
safely after many adventures. He hung the sun in its 
basket in the far east, then surveyed it. It did not light up 
to suit him, and he moved it a little higher. Still it did 
not suit him, so he continued to move it, on and on. And 
he is moving it to the present day." Thus the Indian ac- 
counts for the moving of the sun, and thinks not that the 
earth moves. 

As knowledge has gradually been accumulated regard- 
ing our central luminary, and as information is secured 
about the laws, physical and chemical, to which the sun is 
subjected, the more and more difficult has become the prob- 
lem of finding an explanation adequate to satisfy all of the 
facts. And now in the twentieth century, it is discovered that 
each and every one of the countless billions of chemical atoms 


that form the sun is a solar system in miniature, with the 
result that solar theories must be revised and hypotheses 
revamped in order to take account of electrons and protons. 
The sun is a typical star, but owing to its proximity it may 
be examined in detail, its surface, the spots that are of 
such great interest, the reversing layer, the chromosphere 
and the far-flung corona. The stars are so far distant that 
they appear practically as points of light even in our larg- 
est and best telescopes, and consequently little more can 
be learned of them than what depends on their surface 
brightness. On account of the closeness of the sun, its sur- 
face may be covered up by the interposing moon, and as 
a result, envelopes of chromosphere and corona are shown. 
Although similar envelopes unquestionably exist on the stars, 
they can never be made manifest to us. By the study of 
eclipses a wealth of knowledge is acquired concerning the 
sun, but to gain an adequate idea of what additional in- 
formation is thus secured, it will be well to give a brief 
resume of the salient points of solar research. 

Compared with terrestrial standards, the distance to the 
sun is colossal and its diameter enormous. The problem of 
finding the distance of the sun is one of the most important 
as well as one of the most difficult in the whole of the 
science of astronomy. The importance lies in the fact that 
the distance from earth to sun is the unit for measuring all 
celestial distances, except that to the moon, the solar dis- 
tance being called the astronomical unit. The yard, or the 
meter, is the standard of length for all measurements in 
civilized countries. The " standard yard/' or the " standard 
meter," is a bar of certain composition, of definite 
shape, whose length can be determined by precise measure- 
ments at known temperatures. The standards are kept in 
London or Paris, but various prototypes are widely dis- 
tributed. Certain advantages would result from the em- 
ployment of but one standard of length, the meter, which 
is now almost exclusively used in all scientific measure- 
ments. The reason is not because the length of the meter 
is more valuable than that of the yard, but rather that the 
decimal system employed with the meter is simpler than the 

Two spectroheliograms showing unusual solar activity. Yerkes Observatory. 








more cumbrous division of the yard. At the end of the 
eighteenth century the meter was designed by the French 
to represent the ten-millionth part of the quadrant of the 
earth, so that its circumference should be exactly forty 
million meters. At that time, however, the size of the earth 
was known with little accuracy. If, therefore, the meter 
were actually to represent a certain definite and fixed 
fraction of the earth's size, it would be impossible to use 
it as a standard for the reason that its length would change 
with every revision of the earth's measurement as new 
geodetic operations were carried out, and would alter with 
any variations in the earth itself, which changes geodesy 
and geology tell us are continually taking place. 

By means of the careful researches of mathematical as- 
tronomy stretching back over hundreds, and even thousands 
of years, an accurate plot can be drawn to scale of the 
orbits of all the planets and satellites of the solar system. 
The periods of each have been determined accurately, the 
shapes of their orbits, their inclinations to the ecliptic, etc. 
All of the planetary distances have been found by the as- 
tronomer by referring them to the astronomical unit, the 
distance from earth to sun. To know the complete scale 
of the astronomical plan in miles, it is necessary to know 
accurately at least one distance, either that of the earth to 
the sun, or to one of the planets. Manifestly, the nearer 
the planet, the more accurately can the distance from the 
earth be determined. Celestial distances are usually found 
by the methods of the surveyor who wishes to ascertain the 
width of a river which he cannot traverse. A base-line is 
measured with as great precision as possible, and from each 
end of this base-line, angles are measured to some well- 
defined object on the other side of the river. A civil 
engineer would be in a quandary if a base-line of only three 
inches in length was available for measurement on his side 
of the river and it was necessary to determine, with the pre- 
cision necessary for planning a cantilever bridge, the dis- 
tance across the river. This is exactly the problem that con- 
fronts the astronomer in attempting to measure directly the 
distance of the sun, since the only base-line available is a 


chord of the earth. Instead of expressing the distance to 
the sun in miles or kilometers, the astronomer knows this 
unit ordinarily by means of the small angle at the sun sub- 
tended by the earth's radius, and this angle is called the 
" solar parallax." As there are many different terrestrial 
radii, the earth not being a sphere, the equatorial radius is 
assumed. The astronomical unit when expressed in miles 
must therefore be subject, not only to all of the errors of the 
astronomer in carrying out his measurements, but also to 
those due to the work of the geodesist in determining the 
shape and size of the earth and in referring these measures 
to the standard yard or meter. 

Although this is not the place to discuss the details of 
the determination of the solar parallax, brief references 
will be made to the more promising methods. It is im- 
possible to secure the distance to the sun by direct measures 
and recourse must be had to the determination of this dis- 
tance indirectly by the measurement of other distances in 
the solar system. Great were the expectations aroused by 
the transits of Venus in 1761 and 1769, and again in 1874 
and 1882. By the year 1882, the dry plate had been 
invented and it was anticipated that photography would 
revolutionize our knowledge of the fundamental unit. But 
alas! the " black drop," and the atmosphere of Venus 
caused the observations to be practically a dismal failure. 
Another attempt by this method will not be possible until 
the year 2004, the year of the next transit of Venus. 

Great advances in precision were made by Gill in his 
measures of Mars and some of the minor planets by means 
of the heliometer. The discovery in 1898 of the planetoid 
Eros, which at perihelion comes closer to the earth even 
than Mars, gave to the astronomer a splendid opportunity 
of determining the solar parallax for the reason that 
photography could be applied to the problem. Although 
much was accomplished at the opposition of 1900-01, when 
Eros at its nearest approach was 30,000,000 miles from 
the earth, more will be effected from observations made in 
1931 when the planet was within half this distance. 

Methods based on the law of gravitation furnish the 


means of determining the distance of the sun. E. W. 
Brown's magnificent investigation of the motion of the 
moon gives 8."778 as the value of the solar parallax. Other 
gravitational methods furnish the mean value of 8. "780. 

The velocity of light can be utilized in three different 
manners: (a) by the constant of aberration; (b) by the 
eclipse of Jupiter's satellites; and (c) by the velocity of 
the earth in its orbit by utilizing the Doppler principle of 
measuring the motion in the line of sight from spectra of 
stars or of planets. St. John and Nicholson at Mount 
Wilson, using large dispersion on the spectrum of Venus, 
find the solar parallax of 8."8i3. (Publications A. S. P., 
32, 332, 1920.) They give also the values by other ob- 

A summary of the best values * of the solar parallax (see 
Abbot, The Sun} are: 

From heliometer work on minor planets 8."8o7 

From the Eros campaign 8. 807 

From all gravitational methods 8. 780 

From the eclipses of Jupiter's satellites 8. 799 

From the velocity of light and constant of aberration (2o"47) 8. 803 

The present accepted value of the solar parallax is 8. "80, 
corresponding to a distance of the sun of 92,900,000 miles, 
with an uncertainty of about 50,000 miles, or an error of 
about one part in 2000. To walk the distance to the sun at 
four miles per hour and ten hours per day, 68 years would 
be necessary for the first million miles, or 6300 years for the 
total distance. An express train going at sixty miles per 
hour would take 175 years, while light which travels at the 
great speed of 186,300 miles per second takes 499 seconds. 

If we know the angular diameter of the sun we can readily 
find its linear diameter of 865,000 miles, which is 109.5 
times the diameter of the earth. Perhaps the best method 
of visualizing the huge size of the sun is to compare it with 
the distance to the moon, which in round numbers is 239,000 
miles. If the earth could be placed at the center of the sun 
and the moon were allowed to revolve in her orbit about the 

1 See also, Newcomb-Engelmann, Populdre Astronomic. 


earth there would be plenty of room inside the sun for the 
moon to make her monthly journey, since the moon would 
be little more than half way out to the sun's surface. A 
spot on the sun having a diameter of 8000 miles, or the size 
of the earth, would be regarded as fairly small and a tele- 
scope would be needed to detect it. Dividing the linear 
diameter of 865,000 miles by the angular diameter 1920", 
it is found that i" at the sun corresponds to 450 miles, which 
is equivalent to 725 kilometers. These are useful quantities 
to remember, especially when considering the subject of 

Since the surfaces of spheres are proportional to the 
squares, and the volumes proportional to the cubes of their 
radii, it is readily found, by squaring and cubing 109.5, that 
the surface of the sun is 12,000 times that of the earth while 
its volume is 1,300,000 times the volume of the earth. 

The mass of the sun is 332,000 times the mass of the 
earth, a relation that can be determined by comparing the 
distance a body falls towards the earth in a second of time 
in obedience to the law of gravitation with the distance that 
the earth falls towards the sun in the same interval of time 
and also in obedience to the law of gravitation. In one 
second the earth travels eighteen and a half miles of her 
annual journey about the sun, but as this is accomplished 
without friction we feel no sensation from this rapid flight. 
In going eighteen and a half miles, the earth deviates but 
one-ninth of an inch from a straight line. As the earth 
weighs six thousands of millions of millions of millions of 
tons, which is 6 x io 21 tons, the sun weighs 2 x id 27 tons. 
This is such a colossal number that it makes little differ- 
ence in our comprehension of it whether it is the American 
ton of 2000 pounds, or the English ton of 2240 pounds, or 
the long ton that we pay for when we buy coal or the short 
ton that is furnished by the dealer when the coal is placed 
in the bin. 

Knowing the mass and volume of the sun compared with 
that of the earth, we find the density of the sun is 0.255 times 
that of the earth, or about 1.4 times the density of water. 
The attraction of gravity at the surface of the sun is 27.6 


times that which it is at the earth's surface, so that a man 
weighing 150 pounds would weigh over two tons if trans- 
ported to the sun, and his feet would be so heavy, even if 
the footing were secure, that he would not have strength 
sufficient to lift them. 

The small density of the sun, being only one-quarter that 
of the earth, is one of the most significant bits of knowledge 
connected with the study of the sun. All theories of evolu- 
tion point to the fact that the sun and earth are made of the 
same materials. Indeed Rowland was wont to say that if 
the earth were heated to incandescence it would give a spec- 
trum identical with that of the sun. The low density makes 
it evident that the sun cannot be a solid like the earth, nor 
indeed can it be a liquid, and it must therefore be a gas, the 
terrific heat of the sun being sufficient to vaporize all known 
terrestrial substances. The condition of immense heat and 
enormous pressure caused by gravitation on the sun cannot 
be even distantly approximated in our laboratory experi- 
ments. Unquestionably the sun does not obey the labora- 
tory laws to which such perfect gases as oxygen, nitrogen 
and hydrogen are subjected. For the complete explanation 
of solar phenomena it is necessary to proceed from known 
conditions to those impossible to duplicate in the laboratory 
by the difficult and uncertain methods of extrapolation. 
The steps of scientific development must accordingly be 
carefully planned and wisely thought out, or else the path 
of truth may lead away from the goal of progress rather 
than towards it. 

The spherical portion of the sun that we see is called the 
photosphere. According to Young, The Sun, page 109, the 
" photosphere is a sheet of self-luminous cloud; possibly like 
the clouds of our own atmosphere, with the exception that 
the droplets of water which constitute terrestrial clouds are 
replaced in the sun by drops of molten metal, and that the 
solar atmosphere in which they float is the flame of a burn- 
ing fiery furnace, raging with a fury and an intensity beyond 
all human conception." This notion of the photosphere 
propounded a third of a century ago has been greatly 
modified by modern research. By means of convection cur- 


rents, gases from the interior of the sun are brought to the 
surface. There set free from the enormous internal pressure 
and meeting the cooler temperatures of outside space the 
gases expand. The pent-up energy being suddenly released, 
there is a rapid fall of temperature, and according to Young's 
theory, small solid or liquid particles are formed. By gravity 
these sink back to the solar furnace, there to be changed 
again to the gaseous form. The rising and falling back 
again to the solar surface of the " drops of molten metal " 
cause a continued rain of meteors on the sun's surface. 
Two decades ago these " drops " played a very important 
role in many solar theories, but particularly in that regard- 
ing radiation pressure (Chapter XXI). The temperature of 
the photosphere " seems to be certainly in excess of 6000 
absolute, Centigrade. There are no substances, so far as 
known, which can exist except as vapors in these conditions. 
Hence, it seems reasonable to suppose that the sun contains 
no solids or liquids, unless perhaps in sun-spots, and that its 
substance, as we see it, and within the layers we see, is 
altogether gaseous." l 

In carrying out investigations regarding the sun's surface 
there are two points of view that astronomers should never 
forget. The first is that the photosphere can be viewed or 
photographed only through the superposed layers of the 
solar atmosphere. The photospheric spectrum which must 
be continuous from red to violet without breaks, can never 
be obtained. The second point to be remembered may be 
visualized by analogy with the earth. At an elevation of 
three and a half miles above sea-level, atmospheric air has 
its density cut in half. But gravity at the sun is nearly 
twenty-eight times its value on the surface of the earth. 
Allowance being made for the hundred-fold diameter of the 
sun when compared with that of the earth, it is seen that 
within ten or twelve miles of the sun's surface there would 
exist one-half of the total material in the sun's various layers 
of gases were it not for the enormous temperature of the 
sun. The decrease in pressure upwards from the sun's 
surface is consequently extremely rapid. The importance of 

1 Abbot, The Sun, p. 243, 1911. 


this point cannot be over-emphasized, since most investi- 
gators seem to forget that the change in pressure must take 
place at such a very accelerated rate. 

The sun may be viewed with a telescope by the use of 
solar eye-pieces of various kinds; or by projecting the sun 
on a screen, the solar image being brought to a focus by 
slightly drawing out the telescopic ocular. If the telescope 
is of moderate size, and the definition good, the surface of 
the sun looks like " rough drawing paper, or like curdled 
milk seen from a little distance." The use of a large tele- 
scope and of moments of exquisite seeing that come but 
rarely reveal an infinite wealth of detail. The best drawings 
of the sun are by Langley who describes the surface of the 
sun as that of " snow-flakes sprinkled sparsely over a grayish 
cloth." Before the application of photography great was 
the diversity of opinion concerning the ultimate nature of 
the light-giving particles of the sun. We learned then of 
" rice grains," of " willow leaves," of " thatch-straw " and 
of " granules "; and the various camps in favor of one or 
other designation were about equally divided. 

Photographs of the sun may be obtained by the method 
known to the great host of camera users, that of the focal- 
plane shutter. This consists essentially of a slit of variable 
size that can be driven by means of a spring across in front 
of and close to the photographic plate. By regulating the 
width of the slit opening and the tension of the spring, 
exposures may be varied at will. The exposures necessary 
to obtain good photographs of the sun depend mainly 
on the size of the telescope, the sensitivity of the 
photographic plate and the method of development. The 
modern dry plate of great rapidity does not permit the se- 
curing of solar photographs of the greatest detail. With the 
sun, where there is such an abundance of light, it is unneces^ 
sary to make use of the fastest plates which are primarily fot 
the purpose of decreasing the exposures. Better results may 
be secured by the use of finer grained, slower and more con- 
trasty plates. As a matter of fact, the modern dry plate 
cannot furnish the exquisite definition secured by the old 
wet-plate process, and for this reason the superb solar photo- 


graphs of Janssen at Meudon are unsurpassed even at the 
present day. 

The best conditions for observing the sun are found not 
more than one per cent of the time spent at the telescope. 
Under these maximum conditions, the skilled eye can see 
finer details than can be portrayed on the photographic 
plate. According to Langley, the " snow-flakes " are in the 
neighborhood of 50 to 100 miles in diameter, and these in 
turn are made of flakes similar in form, but of one-fifth the 
dimensions. These small particles cover but one-fifth of 
the surface but radiate three-quarters of the total solar light, 
and hence they must shine with an intensity twenty-fold 
that of the darker portions of the sun. 

The most noticeable feature of photographs showing the 
whole solar disk is the darkening that is found near the edge 
of the sun. This darkening is caused by the absorption by 
the sun's atmosphere, a beam from the limb of the sun pass- 
ing through a greater layer than one from the center. The 
sun or moon when rising or setting looks reddish to us on 
account of the absorption of the blue and violet by our 
terrestrial atmosphere. In a similar manner, the solar atmos- 
phere absorbs more and more of the violet end of the spec- 
trum as the limb of the sun is approached, and the maximum 
of radiation is displaced towards the red. A similar shifting 
of the wave-length maximum is found when comparing the 
spectrum of a sun-spot with that of the photosphere. Abbot 
(The Sun, page 107) gives measures of the distribution of 
radiation over the sun's disk from the center outwards to 
the edge. If the sun could be viewed without the absorptive 
effects of its own and the earth's atmospheres, the maximum 
intensity in the spectrum would be shifted by an appreciable 
amount to the violet, and the sun instead of appearing as 
yellow in color to the eye would look bluish. 

The exquisite photographs of Janssen show the solar 
granulation in splendid detail, the features being sharp and 
well-defined. Other parts of the same photographs are quite 
smudgy in comparison, as if the solar surface were in violent 
commotion. To these parts Janssen gave the name reseau 
photospherique. If photographs taken in rapid succession 


" Like snowflakes sprinkled over a grayish cloth." 
Photographed by Janssen at Meudon, September 9, 1883. 


are examined it is found that the smudgy and ill-defined por- 
tions exist at different parts of the solar image. The 
simplest and most apparent explanation seemed to be that 
these changes afforded positive evidence of violent commo- 
tion on the sun which certainly must exist there on account 
of the very high temperature. This evident explanation, 
however, seems not to be the true one. Any motion of the 
atmosphere of the sun, or of the earth's atmosphere close to 
or far away from the photographic plate would have the 
effect of blurring the photographic image. To make a long 
story short the general opinion regarding the reseau 
photospherique is that it is not a solar phenomenon at all, 
but is caused by the disturbance of the air heated in the 
telescopic tube by the sun's beams. The portions of the 
photograph in good definition represent the true granulation 
of the solar surface. Direct photographs similar to those of 
Janssen have been made by Hansky of Poulkova and 
Chevalier of the Zo-Se Observatory in China. Exposures 
made in rapid succession give the following information re- 
garding the ultimate nature of the photosphere as depicted 
by photographs: (i), The solar granules have a diameter 
of 400 to 1 200 miles, though at times smaller granules are 
seen no more than 100 miles in diameter. (2), They are 
generally circular, or elliptical in shape. (3) , They coalesce 
to form larger particles. (4), These granules are the 
" clouds " of Young's theory. (5), The life of one of these 
clouds is very short, the majority of them last for approxi- 
mately half a minute, and practically none exist longer 
than a few minutes. (6), The displacements vary widely 
in direction and in velocity. The movements range from 
zero to thirty kilometers per second, though occasionally 
higher speeds are observed. (7), The granules in fact seem 
to be the summits of a fleecy structure of condensed par- 
ticles. In fact, they represent l on an enormous scale a 
phenomenon similar in appearance to a storm-tossed and 
choppy sea when viewed aloft from an airplane. 

The most prominent features of the solar surface are the 
spots. Individual records of these exist as far back as the 

1 Astrophysical Journal, 27, 12, 1908. 


Chinese, but their real history begins with the invention of 
the telescope; and they were independently discovered by 
Galileo, Fabricius and Scheiner. There is a great wealth of 
scientific literature connected with the study of spots but 
here it will be possible to give only the salient features, and 
a brief summary of our present knowledge, which, alas! is 
far from complete. As in the study of the photosphere, the 
details of the appearance of spots can be better observed 
visually than by photography. A normal spot consists of 
an umbra, more or less round, surrounded by a less darkened 
penumbra, the structure of the constituent parts of a spot 
differing much from each other and from the surface of the 
photosphere. The roundish shapes of the granular photo- 
sphere are changed in appearance to the straw-thatch of the 
penumbral filaments which exhibit a great wealth of detail, 
and these in turn transform into the smooth, black, velvet- 
like appearance of the umbra. The umbra is, however, not 
uniformly black but is more or less cloudy in appearance 
when conditions of seeing are at the best. Generally asso- 
ciated with spots are the jaculae, or bright patches on the 
sun. These exist at slight elevations above the average sur- 
face of the sun, and are best seen when near the edge of the 
sun where the greater absorption of the sun's atmosphere 
and the elevation of the faculae make them visible by con- 
trast with the darker surroundings. The umbra of a spot 
is dark only by contrast with the more dazzling photosphere, 
yet withal it is not black, for it is more brilliant than the 
electric arc. During the progress of an eclipse of the sun 
a spot has been observed by Evershed to be much brighter 
than the dark limb of the moon occulting it. Langley esti- 
mates that the blackest spot gives 500 times as much light 
as an equal area of the full moon. 

Spots vary in size, from a few hundred miles to 50,000 
miles in diameter in the case of the very largest spots. 
Groups of spots may extend across one-sixth of the diameter 
of the sun, and consequently may be visible to the naked eye 
when the sun is seen through haze or near the horizon, or 
when the eyes are protected by smoked glass. Spots have 
usually a short life, sometimes disappearing in a day or two, 


sometimes lasting for a month or longer. The longest record 
is that of the spot seen during the years 1840-41 which per- 
sisted for eighteen months. Owing to the violent solar mo- 
tion the changes in sun-spots are naturally very rapid, the 
disintegration of spots taking place usually by the formation 
of a bright " bridge " which may be shot across a spot at a 
high rate of speed (compared with terrestrial motions) , of 
as much as one thousand miles per hour. The elevation of 
sun-spots with respect to the general photospheric level is 
still being actively discussed, even after the lapse of a cen- 
tury and a half since 1769, when Dr. A. Wilson of Glasgow 
first propounded his well-known theory, that the fore- 
shortening of the penumbral filaments as the spot neared 
the sun's edge showed that the spots were saucer-like de- 
pressions in the sun's general surface. Spots having been 
seen which confirmed the theory while others seemed to dis- 
prove it, the arguments have gone on pro and con. For a 
more complete discussion, see Agnes M. Clerke, Problems in 
Astrophysics. In view of the very rapid change in gravita- 
tion near the surface of the sun already noted, it is alto- 
gether probable that the elevation at which sun-spots exist 
differs less than fifty kilometers from that of the photo- 
sphere. This small allowable difference of altitude is not 
sufficient to permit the saucer-like sinks needed for Wilson's 

The most evident fact concerning the spots is their peri- 
odicity, first discovered in 1843 by Schwabe, the average 
period being 11.13 years. There are marked differences in 
the length from maximum to maximum, and equally great 
divergences in the intensity of the various maxima so that 
it must be said that spots are very irregular in their regu- 
larity. The individual periods range between 7.3 and 17.1 
years as extremes. But no matter how divergent the period 
is from the mean, the rise to maximum spottedness always 
consumes less time than the descent to minimum. On the 
average the intervals are 4.62 years and 6.51 years, respec- 
tively. The sun-spot curve thus resembles the light curve 
of the average variable star of long period, and also those 
of the Cepheids. The importance of this fact is here 


emphasized for the reason that we have learned from the 
researches of Abbot that the total radiation of the sun varies 
in amount, and as a consequence the sun must be regarded 
as a variable star of long period. 

Various attempts have been made to examine the sun-spot 
curve by the methods of harmonic analysis in order to find 
any secondary periods that may underlie the main period 
of 11.13 years. The most notable attempts in recent years 
have been by Schuster/ Hirayama, 2 Kimura, 3 Michelson, 4 
and Larmor and Yamaga. 5 Kimura and Michelson each 
examined 160 years of sun-spot records from 1750 to 1910, 
and although the material for examination was the same, the 
conclusions reached are greatly at variance. The former 
decides, " The n-year period is not so conspicuous as gen- 
erally considered. Although the most important of all, yet 
to my surprise there are a great many periodicities lying be- 
tween 8 and 12 years, most of them being of considerable 
relative amplitude. " Kimura predicts the form of the sun- 
spot curve up to the year 1950, the predictions giving a maxi- 
mum of spots at the beginning of the year 1914, with an 
intensity about equivalent to that of the maximum of 
iqos-6. According to observations at Mt. Wilson, however, 
Nicholson G finds that the maximum did not take place 
until August, 1917, over three years later than the pre- 
dicted time, while the activity was considerably greater 
than at the preceding maximum. Michelson concludes 
that "with the exception of the n-year period and pos- 
sibly a very long period (of the order of 100 years), 
the many periods found by previous investigations are illu- 
sory. " Quite similar deductions are drawn by Larmor and 
Yamaga who even go so far as to state that when the peri- 
odic part is removed the residue of the sun-spot activity is 
of a fortuitous sporadic character, not amenable to further 
analysis. Various attempts have likewise been made to ex- 
plain the sun-spots by means of the attractions of the 

1 Phil. Trans. Roy. Soc. 206. 

2 Tokyo Sugata, 3, g. 

3 M. N. R. A. S., 73, 543- 

4 Astro physical Journal, 38, 268. 

5 Royal Soc. Proc. A., 93, 493. 
Publ. A. S. P., Ji, 223, 1919. 


planets, particularly those of the giant of the sun's family. 
But Jupiter's period is n.86 years, and even when other 
periods of other planets are superimposed, no success has 
followed the attempts to explain sun-spots by planetary 
influences. The cause of the cycle indeed seems as much 
unknown today as when a hundred years ago Schwabe first 
began his systematic observations. 

Spots are never found more than 45 from the equator 
and seldom at the equator. A curious distribution of the 
spots in latitude manifests itself during the progress of the 
solar cycle. Approximately at the time of minimum, sun- 
spots begin to manifest themselves in two zones, more than 
30 north and south of the equator. With the lapse of time, 
the spots are found closer and closer to the equator, the 
sun-spot maximum taking place with the spots in zones 17 
north and south. The disturbance gradually dies out in 
latitude 8 or 10, after a lapse of 13 or 14 years from the 
first outbreak. Before the final flickering out, the new cycle 
has begun to manifest itself, so that near sun-spot minimum 
there are found four zones of disturbance, two near the 
equator, and two farther north and south. 

Many serious attempts have been made to connect sun- 
spots with various phenomena, solar and also terrestrial. 
The correlation with the solar manifestations mainly rests 
on secure ground, but with some of the earthly influences, 
the connection seems rather far-fetched. If it is hotter than 
the average at a certain locality in the United States, like 
St. Louis, or if mayhap at the same time, Northern France is 
having a cold spell, and if coincidently there is a large spot- 
group on the sun, an astronomer, or usually pseudo-astrono- 
mer, is always found who informs the daily press that the 
sun-spot is the cause of the heat (or cold). The famines in 
India, the potato crop in Ireland, the price of corn in Eng- 
land, the rain- fall in the Island of Mauritius, the financial 
panics of Wall Street all have been investigated by statistical 
methods, and each and all have been found to pass through 
periods of the same length, and to be connected with the sun- 
spot period. It has many times been said that " figures never 
lie." It is quite true that the figures themselves do not tell 


falsehoods, but many and varied are the interpretations that 
may be placed on these figures. The president of every big 
business corporation well knows that if his company shows 
in two successive years the same approximate amount of 
earnings, it is very easy for him to declare a substantial 
dividend, or to charge certain amounts to " improvements/' 
and curtail or even pass a dividend. The whole question 
depends on whether he wishes to please the fifty-one percent 
of the majority stock-holders (of whom he is one), or to 
take advantage of the forty-nine percent (of whom he is not 
one). It is quite possible, and indeed probable, that the 
weather and rainfall are connected with variations of solar 
activity as evidenced by sun-spots, and that other manifesta- 
tions of meteorological changes are also correlated, but to 
prove the connection " is another story." The weather is 
" not made on the spot/' it depends on a vast variety of 
conditions and it is therefore difficult and well-nigh im- 
possible to single out the solar cause from all of the possible 
influences that may affect the weather. It might not be out 
of place to remind all such investigators that a great variety 
of periods have been found in the sun-spot cycle itself, but 
that none of the subsidiary periods appear to have any 
reality, no matter how firmly substantiated by figures they 
seem to be. Most of the meteorological dependences seem 
to be equally illusory. There are, however, many terrestrial 
and many solar phenomena which have a well proven con- 
nection with sun-spots. These will be given below in brief 
form, though some of them will later be expanded more 

Records kept at the Greenwich Observatory and extend- 
ing over nearly a hundred years, show (i), that the diurnal 
range of the magnetic declination, and (2), that the hori- 
zontal force of the magnetism flowing through the earth, 
follow the sun-spot fluctuations not only in the main n-year 
period but even in the small and secondary variations. The 
parallelism is so intimate that it is at once evident that if 
the cause can be found of the sun-spot cycle there also will 
be found the true explanation of the variation of terrestrial 
magnetism. Although the records are not so complete, (3), 


aurorae and (4), magnetic storms are more abundant when 
spots are numerous. The same may be said (5), of faculae, 
and (6), of prominences. (7) The shape of the corona 
changes with the sun-spot period; at minimum of spots there 
are long equatorial extensions, and well defined polar rays. 
(8) The conclusions of Koppen, Stone, Gould, Nordmann, 
Newcomb, Abbot and Fowle, Arctowski and Bigelow are 
that there is a change in the mean temperature of the earth, 
small in size, but amounting to 0.7 between sun-spot maxi- 
mum and minimum, the earth being cooler at sun-spot 
maximum, which, in fact, is quite contrary to the ordinary 
popular belief. 

The contribution of the spectroscope regarding spots and 
related phenomena will be given in a subsequent chapter. 

The state of our knowledge of the angular diameter of 
the sun is far from satisfactory. The accepted value of this 
fundamental quantity comes from measurements made many 
years ago with the heliometer. The discussion by Schurr 
and Ambronn gave the angular diameter of the sun when 
at its mean distance from the earth to be equal to 1 920^.0 
=*= o".c>3. The investigations showed that there were slight 
but unmistakable differences between the polar and equa- 
torial diameters, but in spite of these differences it was as- 
sumed that the sun was spherical. In this procedure they 
followed the example set by the great master, Auwers. 

In the present day of refinement in solar research it seems 
quite unsafe to assume that the sun is necessarily spherical 
or that the changes taking place in the polar and equatorial 
diameters are so small that they are beyond the possibility 
of measurement. The heliometer, as the name signifies, was 
invented for the purpose of measuring the diameter of the 
sun, the first heliometer coming from the hands of Fraun- 
hofer (p. 77). It is a very valuable and refined instrument 
of measurement, and in addition to its use on the sun it has 
been extensively employed in the determination of stellar 
parallaxes. The latter research has shown its limitations. 
When used on the sun, the great heat of the sun causes the 
same effect which is found in every instrument of precision 
when exposed to the sun's rays, namely, the instrument is 


put out of accurate adjustment. The heliometer being 
specially sensitive to changes in adjustment it is not sur- 
prising to find, by reference to the original observations, 
that the same observer, using the same instrument on two 
successive days and under good observing conditions, would 
obtain differences in the measurement of the equatorial di- 
ameter of the sun amounting to i", 2", 5", or even 10" or 
more. In addition, every observer has a " personal equa- 
tion " in that he may constantly measure a quantity smaller 
or larger than the average observer. Each astronomer who 
has made visual or photographic measures is familiar with 
the problem of systematic errors. With the sun it is difficult 
to know how to allow for irradiation which causes a spread- 
ing of the image. 

The only practicable method of combining observations 
inconsistent among themselves is to group all the measures 
together and take the mean, at times assigning different 
weights. If the quantity of observations is sufficiently great 
and the number of observers numerous enough, it is quite 
safe to assume that the peculiarities of any one individual 
will have little effect in the final mean. For the determina- 
tion of the diameter of the sun, Auwers had at his disposal 
no less than 15,000 observations made by 100 observers 
working between the years 1851 and 1883. The measures, 
however, were not all made by the heliometer. 

The large systematic errors to which these measures were 
subject have had a curious effect on future observations, a 
parallel to which is not found in any other department of 
astronomical investigation. The result has virtually been to 
terminate all observational measures of the solar diameter. 
What would it avail any astronomer if he should measure the 
diameter of the sun with the heliometer on hundreds of days, 
spending perhaps many thousands of hours of diligent toil 
in the research, only to find that his results differed from the 
generally accepted value? To the astronomical world this 
difference would probably be regarded as a proof, not that 
the accepted value was in error or that the diameter of the 
sun was changing, but rather that the measures of the in- 
dividual though made with the greatest of refinement were 


subject to personal equations. At best the astronomer se- 
cures little reward for the hard toil devoted to his skilled 
researches. Moreover, he is a human being and naturally 
he desires some compensation other than that of advertising 
to the rest of the scientific world that he is a faulty ob- 
server. And this, strange to relate, is the state of affairs in 
the enlightened days of the twentieth century when hun- 
dreds of thousands of dollars are spent each year in solar 
research ! Astronomy appears to be virtually saying that no 
improvements are possible in the work done nearly half a 
century ago by the heliometer and so we shall assume 
that the sun is spherical and without change. 

Are no other methods available? Why not try photog- 
raphy? Surely the great resources of modern astronomy 
can conquer any difficulties ! There are indeed no difficulties 
of any note connected with the photographic processes, for 
excellent photographs of the sun are being secured daily. 
The main difficulty to be overcome is the same one that 
affects heliometer work, namely, the heat of the sun. This 
alters the length of the telescope tube and changes the focal 
length of the object glass so that the exact scale of the photo- 
graphs is uncertain. These changes, however, do not alter 
the relative scales of the polar and equatorial diameters. 
Some indefatigable worker, therefore, has already waiting 
to his hand some hundreds of thousands of solar negatives 
to be measured and discussed for the purpose of determining 
whether or not the sun is spherical. 

The only method apparently available of eliminating the 
effect of the heat of the sun and at the same time applying 
photography is clearly outlined by Hayn 1 who applied it 
with great success at the eclipses of April 17, 1912, and 
August 21, 1914. By means of photographs taken at the 
time of a solar eclipse, not however, during totality, but 
during the partial phases, the shape and size of the sun can 
be determined, the shape and size of the moon also, and in 
addition, the times of contacts of the limbs of the sun and 
moon usually secured at eclipses. 

A research somewhat similar in character has been carried 

1 Astronomische Nachrichten, 201, 185, 1915. 


out by Henry Norris Russell * who measured photographs 
taken at Harvard College Observatory for the purpose of 
determining the position of the moon with respect to the 
stars. To secure satisfactory photographs, it was necessary 
to make the exposures on the moon the thousandth part of 
those required for the stars, and at the same time the tele- 
scopic object-glass had to be shielded from the light of the 
moon so that the photographic plate might not be fogged. 
The difficulties to be surmounted by Hayn's method during 
the progress of the eclipse are not as great as those overcome 
at Harvard in thus photographing the moon. To give in- 
formation of the highest degree of reliability, it is necessary 
to know the latitude and longitude and the observed times 
with great precision. On this account it would be preferable 
to test Hayn's method at a fixed observatory rather than 
to attempt it under the temporary conditions of an eclipse 
expedition. The best locations will be those nearest the 
path of totality. At the eclipse of September 10, 1923, three 
great American observatories and the Mexican National 
Observatory at Tacubaya were conveniently located, the 
American institutions being Mt. Wilson, Lick and Lowell 
observatories. With the superb instrumental equipment of 
the Mt. Wilson Observatory and an eclipse ninety-eight 
percent total, what magnificent scientific results might not 
have been obtained if the skies had not been densely cloudy! 
At the total eclipse of January 24, 1925, the Van Vleck 
Observatory with its 2o-inch visual refractor was in the path 
of totality, and consequently a series of photographs were 
taken between third and fourth contacts. On account of 
the poor conditions of seeing, everywhere observed at this 
eclipse, the limbs of both sun and moon were not sharp. 
Measures 2 of ten Van Vleck photographs and two made with 
the Yale refractor show the difficulties of making proper al- 
lowance for irradiation, or spreading, of the photographic 
images. It might be possible to allow for uniform irradiation 
but it is more difficult to take care of differential irradiation 
due to the fact that the intensity of the solar radiation de- 

1 Harvard Annals, 72, 76 and 80. 

2 Astronomical Journal, 37, 32, 33, 1926. 


creases toward the limb, thus greatly affecting plates taken 
when the solar cusp is small near the time of totality. Evi- 
dently the only practicable method would be to take photo- 
graphs both before and after totality. 

At the brief eclipse of April 28, 1930, three photographs 
were taken ' with the 36-inch Lick refractor, in spite of wide- 
spread clouds. 

The eclipse of October 21, 1930 gave the opportunity of 
taking sixteen photographs with each of two cameras, one 
38-inches, the other i5-feet in focal length. The lenses were 
fitted with graflex shutters and were stopped down. With 
each camera, eight plates were taken at one-minute intervals 
starting shortly after first contact, and another eight plates 
ending shortly before fourth contact. The plates were used 
to determine the times of first and fourth contacts. Differ- 
ences were found in the intensities of the plates caused by 
variations in transparency and alterations in the rate of the 
exposing shutter. The plates were measured 2 independently 
by Marriott and Pitman. The mean of their measures shows 
that the 1 5-foot camera gave the time of first contact 1.2 sec- 
onds earlier, and the time of fourth contact 1.2 seconds later, 
than was obtained from the 38-inch camera; an effect un- 
questionably due to irradiation. Assuming a correction of 
+ i ".50 to the sun's mean longitude, the photographs showed 
that the American Ephemeris position of the moon needed 
corrections of -j- 5". 54 in mean longitude and 0^.25 in 
latitude. For predicting the eclipse, Washington had used a 
correction to the moon's longitude of + 5"-47 obtained from 
occultations. The photographs showed that first contact was 
1.5 seconds earlier and fourth contact o.i seconds later than 
the predicted times. 

1 Publ. A. S. P., 42, i45, 1930. 

2 Astronomical Journal, 41, 129, 1931. 



"Tycho sought the truth 

From that strange year in boyhood when he heard 
The great eclipse foretold; and, on the day 
Appointed, at the very minute even, 
Beheld the weirdly punctual shadow creep 
Across the sun, bewildering all the birds 
With thoughts of evening." NOYES. 

ASTRONOMY owes much to the eclipse of the sun 
visible in Copenhagen on August 21, 1560, and to 
the fact that a red-headed, freckled-face boy of 
fourteen , destined to become the greatest and most careful 
of observational astronomers since the time of Hipparchus, 
had his keen young imagination fired by watching the 
" orange ember in the sky wane into smouldering ash.' 7 As 
a consequence, this boy, Tycho Brahe, resolved to devote 
his life to unravelling the deep mystery of these strange 
happenings. The romantic incidents of his productive life 
have been beautifully told by Alfred Noyes. From his own 
printing press in his observatory of Uranibourg appeared 
the Historia Coelestis in which appears a long list of eclipses 
beginning with one visible in Rome on March 28, in the 
year 5 A.D. 

At the eclipse of May 3, 1715, Halley referred to that of 
the year 1140 as the last one previously observed in London. 
Although visible not far from London, Hind finds from in- 
vestigation that this eclipse was not seen in the city itself, 
so that it can be said with certainty that not a single total 
eclipse of the sun had visited London for 600 years previous 
to 1715. 

The first eclipse of the sun to be carefully observed in 
the British Colonies of America was that of June 24, 1778, 
which was watched by the astronomer David Rittenhouse 




From the painting by Howard Russell Butler, N. A. 


of Philadelphia. The first American expedition was orga- 
nized and sent out from Harvard College for the eclipse of 
October 27, 1780. As this took place during the war of the 
American Revolution, an appeal was made to " the govern- 
ment of the Commonwealth that a vessel might be prepared 
to convey proper observers to Penobscot-Bay; and that ap- 
plication might be made to the officer who commanded the 
British garrison there, for leave to take a situation con- 
venient for this purpose. 

" Though involved in all the calamities and distresses of 
a severe war, the government discovered all the attention 
and readiness to promote the cause of science, which could 
have been expected in the most peaceable and prosperous 
times; and passed a resolve, directing the Board of War to 
fit out the Lincoln galley to convey me to Penobscot, or any 
other port at the eastward, with such assistants as I should 
judge necessary. 

" Accordingly, I embarked October 9." * 

Probably on account of an error in the tables, the eclipse 
was not total where the Harvard party was located. Be- 
tween the first and second contacts Professor Williams 
measured the angular length of the moon subtended by the 
decreasing crescent of the sun. He gives the following de- 
scription of what appeared shortly before the total phase 
was expected: " The sun's limb became so small as to ap- 
pear like a circular thread or rather like a very fine horn. 
Both the ends lost their acuteness and seemed to break off 
in the form of small drops or stars some of which were round 
and others of an oblong figure. They would separate to a 
small distance, some would appear to run together again 
and then diminish until the whole disappeared." 

Apparently this is a clear description of the so-called 
" Baily's Beads " observed by Francis Baily at the eclipse 
of 1836. An excellent description of this phenomenon is 
given by Agnes Clerke in her History of Astronomy during 
the Nineteenth Century, page 74. Baily gave the correct 
explanation of the phenomenon he saw as being due to ir- 
radiation. This same effect is seen when one holds up his 

1 Memoirs American Academy of Arts and Sciences, /, 84, 1783. 


hand to the sunlight. In making the fingers come close to- 
gether, they appear to touch each other before one feels 
they are actually in contact. An analogous manifestation 
is called the " black drop " which caused surprise at the 
transits of Venus in the years 1761 and 1769, and was the 
source of great trouble to astronomers at the transits of 1874 
and 1882, so widely observed for the purpose of determining 
the solar parallax. The appearance of " Baily's Beads " is 
a phenomenon well worth watching and should be atten- 
tively looked for just before totality begins and just after 
it ends. Very excellent observations may be made with a 
good pair of field glasses or with a small telescope, a large 
telescope being unnecessary. 

Baily was not an astronomer by profession. He was a 
stock-broker, and fortunately he had been successful in the 
making of money, with the result that he was able to devote 
the maturer years of his life to astronomy which he took up 
as his hobby. His work is but one of the many instances 
of the great debt of science to the amateur astronomer. 
One important result of his observations in 1836 was to 
show professional astronomers that at the time of the total 
eclipse of the sun there were other phenomena to observe 
than the mere times of contact of the limbs of the sun and 

The eclipse of 1836 witnessed not only the phenomenon 
of " Baily's Beads " but also an attempt by Forbes to test 
the physical constitution of the sun's atmosphere by means 
of the spectroscope. A new era for astronomy had accord- 
ingly dawned. An eclipse occurred in Southern Europe on 
July 8, 1842, and into the narrow track were collected the 
foremost astronomers from England, France, Germany and 
Russia. What was observed in 1836 was as nothing com- 
pared with the wonders of the eclipse of 1842! 

One of the strangest portions of the history of astronomy 
before the middle of the nineteenth century is the evident 
lack of interest in, or perhaps one should say, the dearth 
of accurate observations of the phenomena visible at 
the time of a total eclipse of the sun. The startling 
suddenness of the apparition, coming in the early days of 


civilization without warning, must have brought terror to 
the hearts of the populace and caused them to fear 
war or pestilence or the death of a favorite prince. 
It is but natural that the prehistoric superstition of 
the dragon swallowing the sun should have spread during 
the middle ages from the far East to all of the civilized 
world. One fleeting glance , however, should have revealed, 
even to the most timorous minded, the pearly-gray light of 
the corona and brought to view the glow of the rosy-hued 
prominences. To those of the present generation, the nine 
hundred and ninety-nine out of a thousand who have never 
had the good fortune to witness a total solar eclipse, it might 
not be out of place to point out that a telescope is entirely 
unnecessary for viewing the beauties of the corona, this 
being a spectacle that derives its glory from the wide-spread 
splendor and slight gradations of contrast. A telescope, 
small or large will of course magnify any particular portion 
but to see and enjoy the beauty of the corona as a whole 
nothing is actually needed but the normal naked eye. 

Published references to the corona in the early literature 
are exceedingly rare. Plutarch and Philostratus give allu- 
sions which unmistakably refer to the corona, but appar- 
ently the first to take any scientific cognizance of the crown 
of glory was Kepler who seems to have witnessed the solar 
eclipse of 1605 in Naples. A hundred years later at the 
eclipse of 1706, Cassini, who was a practised observer, de- 
scribes the " crown " of pale light, and he decides that it 
must be caused by the illumination of zodiacal light; and 
eleven years thereafter, Halley saw the corona and also 
prominences, but he was unable to decide whether the corona 
belonged to the sun or to the moon. 

If so little attention was paid to the corona, it is not sur- 
prising that even less should be taken of the " red flames/' 
though if one refers to the plate facing page 130 he will see 
what a brilliant spectacle they afforded in the eclipse of 
1918. The first reference to them seems to be at the eclipse 
of 1706 when they were apparently observed by Stannyan 
who wrote a description of them to Flamsteed. The first 
vivid portrayal was by Vassinius of Sweden who observed 


them in 1733. The Spanish admiral Ulloa observed them 
while at sea during the eclipse of June 24, 1778, and he 
furnished a valuable account, with the added explanation 
that the rosy hues were caused by the sun's light shining 
through some hole or crevice in the limb of the moon! 

The astronomers who witnessed the eclipse of 1842 were 
entirely unprepared for the phenomena that met their gaze. 
Baily repaired to Pavia, and made his observations from 
one of the rooms of the University. One of the professors, 
out of the goodness of his heart, offered to assist him in any 
way possible, but Baily informed him that all he wanted 
was to be " left alone, being persuaded that nothing is so in- 
jurious to the making of accurate observations, as the intru- 
sion of unnecessary company. " Not being content with this 
gentle hint, the key was taken from the outside of the door 
and it was securely locked on the inside. Baily's report of 
the observations made at the eclipse is found in the Memoirs, 
R. A. S., 15, 4, 1846, as follows: " The beads were distinctly 
visible. ... I was astounded by a tremendous burst of ap- 
plause from the streets below, and at the same moment was 
electrified at the sight of one of the most brilliant and splen- 
did phenomena that can be imagined. For at that instant 
the dark body of the moon was suddenly surrounded with 
a corona, or kind of bright glory. ... I had indeed antici- 
pated a luminous circle round the moon during the time of 
total obscurity, but I did not expect, from any of the ac- 
counts of previous eclipses that I had read, to witness so 
magnificent an exhibition as that which took place. . . . 
The breadth of the corona, measured from the circumference 
of the moon, appeared to me to be nearly equal to half the 
moon's diameter. It had the appearance of brilliant rays. 
Its colour was quite white, not pearl colour, nor yellow, 
nor red. 

" Splendid and astonishing, however, as this remarkable 
phenomenon really was, and although it could not fail to 
call forth the admiration and applause of every beholder, 
yet I must confess that there was at the same time some- 
thing in its singular and wonderful appearance that was 
appalling. . , , But the most remarkable circumstance at- 


tending this phenomenon was the appearance of three large 
protuberances apparently emanating from the circumference 
of the moon, but evidently forming a portion of the corona. 
. . . All of these projections were of the same roseate cast 
of colour, and very distinct from the brilliant vivid white 
light that formed the corona. . . . The whole of these three 
protuberances were visible even to the last moment of total 
obscuration, at least, I never lost sight of them when look- 
ing in that direction; and when the first ray of light was 
admitted from the sun, they vanished with the corona, alto- 
gether, and day-light was instantly restored." 

The same appearance was witnessed by Airy, by Arago 
and others. Arago has an interesting account of the effect 
of the eclipse on the populace who had gathered in great 
numbers to watch the phenomenon. " When the sun, re- 
duced to a very narrow filament, began to throw upon the 
horizon only a very feeble light, a sort of uneasiness seized 
upon all; every one felt a desire to communicate his impres- 
sions to those around him. Hence arose a deep murmur, 
resembling that sent forth by the distant ocean after a 
tempest. The hum of voices increased in intensity as the 
solar crescent grew more slender; at length the crescent 
disappeared and an absolute silence marked this phase of 
the eclipse. The phenomenon in its magnificence had tri- 
umphed over the petulance of youth, over the levity which 
certain persons assume as a sign of superiority, over the 
noisy indifference of which soldiers usually make profession. 
A profound stillness also reigned in the air, the birds had 
ceased to sing." 

The arrival of totality in Milan was greeted by a great 
shout, mingled with cries of " Long live the astronomers " 
who had provided such a beautiful phenomenon to please 
and interest the populace! 

The unexpected nature of prominences and corona seen 
at the eclipse coupled with the publication in 1843 of 
Schwabe's discovery of the periodicity of sun-spots caused 
an unprecedented increase of interest in matters pertaining 
to the physical constitution of the sun. Various ingenious 
explanations appeared. While most astronomers believed 


that the prominences were truly solar in their origin, there 
were many who thought they were possibly some exhalation 
in the earth's upper atmosphere, while still others believed 
that in some manner diffraction round the edge of the moon 
was responsible for these eclipse envelopes. In passing, we 
might mention a notion of no less an authority than Halley 
which was so curious that it should be classed along with 
William HerschePs belief that the sun might be cool and 
habitable. Halley 1 thought that the appearances on the 
eastern and western edges of the sun at a total eclipse might 
reasonably be expected to be different, for the reason that 
" the eastern limb of the moon had been exposed to the sun's 
rays for a fortnight, and as a consequence it would be natural 
to expect that the heated lunar atmosphere might exert some 
absorbing effect on the solar rays, while on the contrary the 
western edge of the moon being in darkness and cold for two 
weeks could exhibit no such absorbing action." 

The interest aroused in total eclipses was now so great 
that astronomers were determined to take advantage of 
every opportunity, no matter how short the time of totality 
nor how great distances it was necessary to travel in order 
to view the eclipses. The eclipse of July 28, 1851, was 
visible in Norway and Sweden, and English astronomy was 
well represented in the persons of the astronomer royal Airy, 
Hind, Dawes, Carrington, Stephenson, Gray, Lassell and 
Williams. Although Faye * still asserted with force that the 
prominences were merely optical illusions or " mirages pro- 
duced near the moon's surface," the general consensus of 
opinion was that the origin of the red flames was to be 
sought in the sun. To this fire of scarlet hue Airy gave the 
name of sierra. 

Any lingering doubts regarding the origin of this sierra 
were forever dispelled by the observations made at the 
eclipse of July 18, 1860, visible in America, Spain and 
Northern Africa. The solution of the problem was accom- 
plished by photography which was applied for the first time 
at an eclipse with anything like success. As the prominences 

1 Phil. Trans., 29, 248, 1715. 

2 Memoirs, R. A. S., 21, 5, 1853. 


are red m color to the eye and as the ordinary photographic 
plate is insensitive to red (plates are usually developed 
under ruby light) grave doubts were felt whether photo- 
graphs would be able to portray the red flames. The only 
thing to do under the circumstances was to " try something 
and see what happens " (excellent advice for the scientist 
usually credited to the late Professor H. A. Rowland). 
Photography had already been applied at the eclipse of 1851 
when Busch obtained some feeble impressions of the eclipsed 
sun by the daguerreotype process. Photography was even 
attempted in 1842 using iodized paper, but with no results. 
Warren de la Rue used the heliograph from Kew, enlarging 
the image before it reached the photographic plate, while 
Father Secchi employed a six-inch refractor without enlarge- 
ment. Photographs of both observers were successful. De 
la Rue was near the Atlantic in Spain while Secchi was on 
the Mediterranean Coast, six minutes of elapsed time being 
necessary for the moon's shadow to travel from one station 
to the other. The conclusions from the 1860 eclipse were: 
i. The prominences are rich in actinic power (now known 
to be due principally to the H and K light of calcium). 2. 
As the moon passed in front of the sun it progressively cov- 
ered and uncovered the prominences, thereby demonstrating 
completely that their origin is strictly solar. 3. During the 
six minutes of elapsed time, changes were noted in some 
prominences but no variations in others. 4. The material 
from which the red flames arise is found around the whole 
solar globe. This is the sierra of Airy, but later, called the 
chromosphere by Lockyer. 

The success attending the eclipse of 1860 came almost at 
the same time with the unraveling of the enigma of the 
spectrum. After two centuries of slow and painstaking 
progress, the crucial experiments had been performed by 
Kirchhoff by means of which the action of the spectro- 
scope was at last understood. As a result, interest in eclipse 
observations was no longer confined to the astronomer alone, 
for many investigations were to be undertaken which were 
not confined by the determination of exact time or position. 
The birth of physical astronomy, or astrophysics, insured 


that henceforth eclipse observations would be of quite as 
much interest to the physicist as to the astronomer, and for 
the interpretation of these observations, research work in the 
laboratory was quite as important as that in the observa- 
tory. What spectrum would the prominences give? The 
answer was not difficult. Apparently the prominences were 
not, as had previously been thought, masses of photo- 
spheric material shot up to great heights by some explo- 
sive or volcanic action on the sun, since it was evident 
that the boomerang-shaped protuberance seen at the eclipse 
of 1851 could hardly have existed under the laws of gravita- 
tion. It was top-heavy and could not have hung there 
above the sun even for the short space of time it was visible 
if it had been composed of the general material forming the 
sun's photosphere. And then there was the distinct differ- 
ence in color noted between the red flames and the body of 
the sun. Manifestly, since all of the gases forming the sun 
did not take part in the solar outburst, the prominences 
probably consisted of a few gases only, perhaps one of their 
chief constituents was the lightest of known gases, hy- 
drogen, whose visible spectrum known from stellar investiga- 
tions consisted of a strong line in the red, another in the blue, 
and a series of others coming closer together as the violet 
end of the spectrum was approached. The red line of hy- 
drogen seemed to give a color not differing materially from 
the red of the prominences. If therefore the prominences 
were actually outbursts of hydrogen gas heated to great 
temperatures in the solar furnace, the eclipse spectrum 
would be vastly different from the ordinary solar spectrum. 
The band of light would not be continuous from the red to 
the violet end, nor would any dark Fraunhofer lines be 
visible. Since the prominences were probably gaseous, their 
spectrum, as known from the laws of Kirchhoff, must consist 
of bright lines on a dark background, an emission spectrum. 
It appeared therefore that the prominence spectrum would 
consist of a few bright lines only, the red and blue lines of 
hydrogen, and the series towards the more refrangible end. 
more difficult to see on account of the fact that the human 
eye is not sensitive to violet light. 


Photographed by Deslandres by spectroheliograph at Meudon, France, on 
June 7, 1919. Above: K/3 of calcium. Below: Ha of hydrogen. Note the 
differences in appearance of the two photographs and the fine detail that 


Photographed by Deslandres at Meudon with the Ki ray of calcium, 

April ii, 1910. 


But there was no eclipse on which to " try and see what 
happened " until August 18, 1868, so it was necessary for 
the physicists and astronomers to possess their souls with 
patience; but alas! the eclipse was visible only in far-off 
India, the Malay peninsula and Siam. The distances were 
great, but the problems were important, and accordingly 
several expeditions, two British, two French, one German 
and one Spanish were found in the eclipse track. The 
greatest success attended the observation of Janssen. The 
slit of his spectroscope directed to the edge of the sun re- 
vealed the spectrum of the prominences consisting, as had 
been thought, of a series of bright emission lines, most 
prominent among which were three lines, the red and the 
blue lines clearly belonging to hydrogen, but one of almost 
equal brilliancy in the yellow. The color of the yellow line 
seemed to match the D-lines of sodium, ever-present in 
laboratory experiments, but why should the gas sodium, 
comparatively heavy, be found in the prominences? 

The brilliancy of the prominence lines was so remark- 
able that Janssen determined to seek them again after the 
eclipse was over. If they were solar in origin they must 
be found on the sun every day, varying it is true in shape 
and dimensions. The only reason why the prominences 
cannot be seen any day is the same reason why the stars 
are not visible in daylight, the glare of the earth's at- 
mosphere, especially when close to the sun, being very great. 
Abolish the atmosphere where the observer worked, and the 
stars and prominences would at once become visible; and 
since the moon has little or no appreciable atmosphere, the 
prominences would be seen each day without an eclipse to 
the " man in the moon " if such a person existed. But how 
get rid of, or diminish, the glare of the earth's atmosphere 
to such an extent that the prominence luminosity will be 
more intensive than the light of the earth's atmosphere? 
This feat can be accomplished by the spectroscope. The 
emission lines, C and F of hydrogen, found in prominences, 
are monochromatic in character, that is, these lines betake 
approximately the nature of mathematical lines and show no 
appreciable width. As a matter of fact, it is impossible to 


diminish their intensity by increasing the dispersion of the 
spectroscope, whether this increase is accomplished by the 
addition of extra prisms or by the employment of a grating. 
Increased dispersion, however, spreads out these lines in 
the spectrum to greater distances apart, but their inten- 
sities are not thereby diminished. On the other hand, the 
continuous background of the solar spectrum can be weak- 
ened at will by merely increasing the amount of the disper- 
sion, a fact which is at once evident since the light 
passing through the slit is by an increase of prisms spread 
over a greater area. Reflected sunlight, whether from 
the moon's surface, from the planets, from a silvered 
mirror or from particles of dust in the earth's atmos- 
phere, gives the solar spectrum. Consequently, the promi- 
nences may be made visible by the spectroscope by the 
simple process of increasing the dispersion to such an 
extent that they can shine by contrast with the weakened 
atmospheric glare. As a result of these ideas, Janssen 
looked for and found the prominences after the eclipse was 
a thing of the past. The same or similar ideas had oc- 
curred to other workers with the spectroscope, notably to 
Huggins and to Lockyer in England. Without having been 
present at the eclipse, the latter tried for the prominences 
and found them for the first time on October 20, 1868. 
Lockyer sent a record of his observations to the French 
Academy, and without having heard of the Englishman's 
results, Janssen sent on to Paris the report of the work he 
had done at the eclipse and afterwards. By a strange coin- 
cidence, the papers from both investigators were read at the 
same sitting of the Academy, in honor of which event a 
medal was struck bearing the likeness of both Lockyer and 

Thus in the moment when the chemical nature of the 
prominences was discovered by the spectroscope, these ob- 
jects ceased to be phenomena confined to eclipses only. As 
a happy and most fortunate result, a two-fold benefit thus 
accrued to the astronomer: freed from the necessity of ob- 
serving prominences during the all too brief moments of a 
total eclipse he could devote his energies to other investiga- 


tions. Not only could prominences be observed without an 
eclipse , but by the same methods, researches could be carried 
out on the solar envelope from which prominences arose, the 
chromosphere. Frequent violent eruptions on the sun 
carried the solar flames up to great distances and with enor- 
mous velocities, and these phenomena were observed visually 
by Young, Lockyer, Tacchini and a host of other investiga- 
tors, many important researches 1 being carried out. The in- 
vention of the spectro-heliograph in 1893, independently and 
almost simultaneously, by Hale and Deslandres permitted 
an attack on these problems by the help of photography. 

The great triumph of the spectroscope in 1868 gave evi- 
dence to the solar astronomer of the important problems 
awaiting solution, and it appeared almost certain that each 
observation carried out with care would be a valuable dis- 
covery. What was the corona? Did it shine by its own 
light, or by reflected sun's light? Was there any connec- 
tion between the prominences and the corona? How far 
out did the corona extend, and were there any changes in 
its form that could be detected? Was the explanation of 
the dark Fraunhofer lines of the solar spectrum the true 
one, and was it possible that eclipses could help in the 

The eclipse of August 7, 1869, crossed America diag- 
onally from Alaska to North Carolina. The United 
States government made a large appropriation to help de- 
fray the expenses, and as a result of this and also on ac- 
count of the favorable time of year, the eclipse track in 
the United States was almost one continuous observatory, 
so thickly were the astronomers scattered. To give a list 
of those who saw the eclipse would be practically equiva- 
lent to making a record of every astronomer of any impor- 
tance who was then living in the United States. Those too 
" were the good old days " before railroad executives were 
harassed by government regulations and Interstate Com- 
merce Commissions and when free transportation for pas- 
sengers and goods could be furnished. The following is 
copied from the Report on Observations of the Total 

i See Young, The Sun. 


Eclipse of the Sun, August 7, 1869. On page 3, J. H. C. 
Coffin, U. S. N., Superintendent of the Nautical Almanac, 
reports, " Col. Thomas A. Scott, vice-president of the Penn- 
sylvania Central Railroad, furnished a special car, and with 
the cordial cooperation of Mr. Robert Harris, general super- 
intendent of the Chicago, Burlington and Quincy, and Mr. 
C. E. Perkins of the Burlington and Missouri Railroad, 
provided free transportation for these parties, with all their 
instruments and apparatus to and from the places of ob- 
servation. Free passes to ten or twelve others over the 
same routes were also granted/' In the same report, page 
115, Professor Morton estimates that " an expense of 
$1500 was spared the government appropriation." 

Fortunately, clear skies greeted the observing parties. 
Little of the important work accomplished will be noted 
in detail here. Spectroscopically, the most valuable dis- 
covery was that the spectrum of the corona was continuous 
but was traversed by a single green ray. This green line 
was detected independently by both Harkness and Young, 
the latter identifying its position as coinciding with the line 
numbered 1474 on Kirchhoffs scale. But since this line 
1474 is due to iron, it was surprising and perplexing in the 
highest degree to find it present in the corona and reaching 
such great heights above the sun's surface. In spite of the 
apparent coincidence, it was evident that the substance 
causing the green line was not iron. To it the name 
coronium was given, and today after more than a half 
century of active research we know little more of coronium 
than when it was first discovered. At this same eclipse, 
Professor E. C. Pickering employed a portrait lens, thus 
recognizing its value in the portrayal of the heavens. As 
he was located at Mt. Pleasant, a little farther west than the 
rest of the expeditions, to him belongs the honor of securing 
the first successful photograph of the corona in America. 
Other photographs were secured by Winlock and others on a 
larger scale and with good definition which showed, the 
corona and prominences. 

The United States Congress appropriated the sum of 
$29,000 for observations at the eclipse of December 22, 


1870, visible in Spain, Northern Africa, Sicily, Greece and 
Turkey. America was represented in the eclipse track by 
the following: Pierce, Newcomb, Harkness, Hall, East- 
man, Winlock, Young, Langley, Pickering, Peters, Watson, 
Clark, Ernst, Willard, Ross, Gannet, and General Abbott. 
The British government granted 2000 and a ship. A 
great variety of observations were projected and only par- 
tially carried out on account of clouds that prevailed almost 
everywhere, observations photographic, spectroscopic, 
photometric and polariscopic. Lockyer met shipwreck and 
clouds, but was rewarded in the end by a glimpse of the 
corona for one brief second and a half! Janssen made 
good his exit from beleaguered Paris in a balloon, but though 
he succeeded in escaping from the bullets of the Germans, 
he was forced to capitulate to obscuration by the clouds. 

The most conspicuous success awaited the efforts of Pro- 
fessor C. A. Young of Princeton who had foretold, and 
whose eye was the first to see the " flash spectrum." Ac- 
cording to the theory of Kirchhoff (p. 97) the spectrum 
of the photosphere would be continuous from red to violet 
without any dark lines were it not for the overlying solar 
atmosphere. Here in the so-called reversing layer, the 
gases are at a lower temperature than in the photo- 
sphere, and on account of these cooler conditions the 
photospheric light is absorbed by the reversing layer, 
and the resultant spectrum is that of dark lines on 
a bright background. As already explained, if the photo- 
sphere could be removed, then the gases forming the revers- 
ing layer, since they are at a high temperature, would give 
a series of bright lines on a dark background, which is 
technically called a " reversal " of the Fraunhofer spec- 
trum. To describe the appearance in 1870, one cannot do 
better than to quote from the words of the discoverer: 1 
" The observation is possible only under peculiar circum- 
stances. At a total eclipse of the sun, at the moment when 
the advancing moon has just covered the sun's disk, the solar 
atmosphere of course projects somewhat at the point where 
the last ray of sunlight has disappeared. If the spectro- 

1 Young, The Sun, p. 83. 


scope be then adjusted with its slit tangent to the sun's 
image at the point of contact, a most beautiful phenomenon 
is seen. As the moon advances, making narrower and nar- 
rower the remaining sickle of the solar disk, the dark lines 
of the spectrum for the most part remain sensibly un- 
changed, though becoming somewhat more intense. A few, 
however, begin to fade out, and some even turn palely 
bright a minute or two before totality begins. But the 
moment the sun is hidden, through the whole length of the 
spectrum, in the red, the green, the violet, the bright lines 
flash out by hundreds and thousands, almost startlingly; as 
suddenly as stars from a bursting rockethead, and as 
evanescent, for the whole thing is over in two or three 
seconds. The layer seems to be only something under a 
thousand miles in thickness, and the moon's motion covers 
it very quickly. 

" The phenomenon, though looked for at the first eclipses 
after solar spectroscopy began to be a science, was missed 
in 1868 and 1869, as the requisite adjustments are delicate, 
and was first actually observed only in 1870." 

The same phenomenon was witnessed by Pye, a member 
of Young's party. The bright lines were so numerous that 
the impression was gained that every one of the thousands 
of Fraunhofer lines was reversed from dark to bright, w r hile 
" the phenomenon was so sudden, so unexpected, and so 
wonderfully beautiful as to force an involuntary exclama- 
tion." 1 Professor Young called this sudden transition the 
flash spectrum. The same phenomenon was witnessed at 
the eclipse of December 12 of the following year by 
Maclear, Herschel, Lockyer and Fyers, the eclipse being 
total in India, Ceylon and Northern Australia (where clouds 
interfered). Again at the annular eclipse of June 6, 1872, 
seen by Pogson in India, and at the total eclipse of April 
1 6, 1874, witnessed by Stone in South Africa, the flash 
spectrum was observed. This is one of the most interest- 
ingly beautiful and important of the phenomena connected 
with total eclipses. During the past half century the flash 
spectrum has been carefully observed at each eclipse. More 

1 Memoirs, R. A. S., 41, 435. 


recently it has been photographed in exquisite detail, and 
has also been photographed at Mt. Wilson without an 
eclipse. (See Chapter XV.) 

Apparently, it had been almost definitely settled that 
the corona is a truly solar appendage, for in addition 
to the spectroscopic evidence in the matter, the photo- 
graphs in 1871 showed the same details in the coronal 
streamers from observing stations widely separated, an 
effect which could not possibly take place if the corona 
was terrestrial in origin. Such then was the state of scien- 
tific information regarding the corona as it existed before 
the eclipse of 1878 which was visible in America. Little was 
known with exactitude, there was very much of perplexity 
and uncertainty and doubt. The unsatisfactory nature of 
the whole subject may perhaps be envisaged in one sentence 
copied from a book by one of the most assiduous of spectro- 
scopic investigators of the time, Norman Lockyer. " Now 
the whole phenomenon of the corona may be defined in two 
words, cool prominences" 

If the knowledge of the corona was disappointing and 
perplexing, the decisions of the spectroscope regarding the 
reversing layer were equally uncertain. Kirchhoffs theory 
demanded that there should be a layer of vapors relatively 
cool surrounding the photosphere. Where was this layer, 
close to or far away from the sun? Was the layer thin, or 
one more extensive? At the eclipse of 1870, Young had 
discovered a flash spectrum apparently of thousands of 
lines suddenly turned from dark to bright. This phe- 
nomenon was visible for only two or three seconds at the 
beginning and ending of totality, and in consequence, the 
reversing layer must be very shallow, approximately 600 
miles in thickness. But were all of the Fraunhofer lines 
turned from dark to bright? Since the appearance was 
visible during the few hurried and excited seconds of an 
eclipse, it was manifest that no exact comparison with 
Fraunhofer lines was possible, nor would be possible until 
the time should arrive when the flash spectrum could be re- 
corded by photography. Moreover some spectral lines were 
seen to turn bright half a minute or more before the be- 


ginning of totality, so that it was evident that not all of the 
reversing layer was confined to the 6oo-mile limit. Still 
other difficulties presented themselves. If the absorption 
did take place in a layer of the solar atmosphere, then the 
light of the photosphere passing tangentially through this 
layer at the sun's limb should experience an absorption 
greater in amount than that from the sun's center. In conse- 
quence, the spectra of the center and limb of the sun were 
compared in great detail, with the resulting discovery that 
any differences in the relative intensities of the lines were 
of such minor character that they could readily be explained 
by the slight darkening of the limb compared with 
that of the center. Kirchhoff believed that the spectral 
lines of any element like sodium were characteristic, 
or in other words, the same element gave always the 
same series of lines no matter how the element was 
vaporized. But when the flame spectrum was com- 
pared with that of the higher temperature of the elec- 
tric arc, and the still greater temperature of the electric 
spark, great differences in the relative intensities of the lines 
were at once observed. Lockyer was the first to call atten- 
tion to, and emphasize the importance of these changes, 
Take the element calcium, for instance. With a Bunsen 
flame and with small dispersion, the chief line visible is in 
the red end of the spectrum. At the temperature of the 
arc, the strongest line is in the blue, at 42 2 7 A. The red 
line is also visible and also two lines in the violet, the H and 
K lines of the solar spectrum. Increase the temperature 
of excitation by using the electric spark, and the two violet 
lines greatly increase their intensities and become much 
stronger than the blue line, while the red line practically 
disappears. Similar changes of intensity were found by 
Lockyer in magnesium, lithium, iron and other elements 
examined, and these conclusions have been abundantly 
verified by all observers since his time. Still further diffi- 
culties presented themselves. When the wave-lengths for 
the various metals were determined, it was discovered that 
different elements had spectral lines with apparently iden- 
tical wave-lengths. It was found that some. of these com- 


mon lines were due to impurities in the metals examined, 
but when these lines were omitted from consideration there 
were still many lines evidently common to two or more 
elements. It was consequently manifest that the identifica- 
tion of lines in the solar spectrum was not the simple opera- 
tion Kirchhoff supposed it to be, and that " the more 
observations were accumulated the more the spectroscopic 
difficulties increased. 7 ' l Although some of the theories and 
conclusions of Lockyer were incorrect, nevertheless spec- 
troscopists are under a great debt to him for the splendid 
series of researches carried out both in the laboratory and 
in the observatory. 

To solve some of these difficulties, Lockyer busily in- 
vestigated the spectra of many elements by his well-known 
method of long and short lines. If an electric arc is ar- 
ranged horizontally and its image is projected on the 
vertical slit of the spectroscope, it is seen at once that the 
lengths of the lines in the spectrum vary considerably. 
Since the core of the vapor between the two carbon poles 
must be much hotter than that on the outside edges, it was 
evident to Lockyer that the short lines were high tempera- 
ture lines that become visible only at the hottest point, that 
of the core, while the long lines were those which could 
exist at different temperatures, at the great heat of the 
center of the arc and at the lesser temperatures of the out- 
side edges. This method of long and short lines thus ap- 
peared to give a ready and convenient means of separating 
the lines of highest temperature, which were comparatively 
few in number, from the balance of the spectrum. 

By arranging the spectra of stars in an orderly series, be- 
ginning with the white stars, and passing through yellow 
stars to those red in color, it was noted by Lockyer that 
there was a steady increase in the total number of lines in 
the spectra, there being few lines in the white stars other 
than the hydrogen series, while the red stars possess an 
enormous number of spectral lines. Thus in the progres- 
sion in the reverse direction, from red stars back to white, 
the spectra show fewer and fewer lines and become simpler 

1 Lockyer, The Chemistry of the Sun, p. 176. 


in character. The same progression from red to white stars 
saw an increase in intensity in the H and K lines due to 
calcium. And as Lockyer was the first to recognize the im- 
portance of temperature in altering the character of spectra, 
so to him likewise belongs the great honor of recognizing 
that changes in stellar spectra show that the white stars are 
hotter than the yellow and these in turn are hotter than the 
red. (Present-day researches show that as the stars increase 
in temperature in the stellar series, the H and K lines do 
not steadily increase in intensity but reach a maximum 
strength, then begin to fade away and are entirely wanting 
in the hottest stars, the blue-white helium stars of spectral 
type B and stars of type O.) 

Momentous in the highest degree therefore were the dis- 
coveries of Lockyer showing the importance of temperature 
in the interpretation of spectra, but his explanations 
of the underlying causes have not borne the weight of time. 
His was the dissociation theory that the chemical atoms 
were continually broken up into elements less complex in 
structure, which exhibited simpler and simpler forms of 
spectra as higher and still higher temperatures were reached. 
As the white stars are those of the highest temperature, and 
as they show practically nothing but the spectrum of hy- 
drogen which is the element of smallest atomic weight, it 
must represent the simplest element with the simplest spec- 
trum. (The most recent researches show that though there 
may be dissociation of the chemical atoms this takes place in 
a manner vastly different from the dissociation demanded by 
Lockyer's theory). 

The sun being but a typical yellow star, this theory can 
be put to the test in attempting to explain the phenomena 
of the solar atmosphere as revealed by eclipses. Lockyer 
assumed " that in the reversing layers of the sun and 
stars various degrees of chemical dissociation are at 
work, which dissociation prevents the coming together 
of the atoms which, at the temperature of the earth 
and at all artificial temperatures attained here, compose the 
metals, the metalloids and the compounds." x In conse- 

1 The Chemistry of the Sun, 201. 


quence, there were grave doubts expressed by Lockye 
whether the chemical elements known from laboratory ex 
periments could at all exist at the great heat of the sur 
except in the cooler parts of its atmosphere. It wa 
imagined by him that this atmosphere consisted of succes 
sive layers " like the skins of an onion/' the layers next the 
sun obviously being the hottest. Hence in the interior layer! 
could exist only " those constituents of the elementary 
bodies which can resist the greater heat of these regions/ 
The spectrum of these inside layers, if such a spectrun 
could be obtained apart from that of the rest of th( 
sun, would therefore not be a reversal of the Fraunhofei 
spectrum. According to Lockyer's hypothesis, the whoh 
of the solar atmosphere is effective in the productior 
of absorption. Young's observation of the flash spec 
trum demanded a shallow reversing layer of a few hun 
dred miles, but Lockyer's theory refused to admit th< 
existence of a reversing layer separate from the super 
incumbent strata. In other words, there could be IK 
division possible into reversing layer, chromosphere anc 
corona; these were but different manifestations of the solai 
atmosphere, the corona being regarded as the outermost am 
cooler parts of an atmosphere having a composite existena 
and obeying the laws of gravitation. In consequence of this 
theory, although Lockyer had himself witnessed the flast 
spectrum at the eclipse of 1871, he at first was forcec 
to doubt, and then actually to deny the existence oJ 
the shallow reversing layer. The flashing out of lines 
that he observed " has been called the reversing layer; 
but I do not now (1881) believe that it is the reversing 
layer for a moment, for, when it comes to be examined 
we shall probably find that scarcely any of the Fraun- 
hofer lines owe their origin to it, and we shall have % 
spectrum which is not a counterpart of the solar spec- 
trum." (Loc. cit.^p. 360.) The solution of the problen 
could not be effected visually during the brief seconds avail- 
able at a total eclipse for observation of the flash spec- 
trum. The interpretation by one astronomer of what was 
observed should be entitled to as much weight as the opin- 


ion of another. There was no hope of a solution of the 
problem until the time should arrive when photography 
could come to the rescue by furnishing a permanent record 
of the flash spectrum which could be compared line by line 
with the Fraunhofer spectrum in order to see whether the 
one spectrum is the exact reversal of the other. 

While these observations, epoch-making in their impor- 
tance, were being made on prominences and reversing layer, 
the corona was not forgotten. Strange as it may now seem, 
there were still many astronomers of repute who believed 
that the origin of the coronal light should be sought, not in 
solar but in lunar or terrestrial causes. There were even two 
theories based on the moon, one that the corona was due to 
the diffraction of solar rays which pass near the moon's edge; 
the other, that the phenomenon was due to reflection of 
solar rays from the irregularities of the moon's surface. 
Another curious theory which found great favor at the time 
was that the corona was due simply to glare in the earth's 
atmosphere. As a result of this hypothesis the corona would 
necessarily be a phenomenon entirely local in its structural 
character, details appearing differently to observers at sepa- 
rate localities. If due simply to atmospheric glare, the 
coronal details should be found projected also on the dark 

There were now available four different methods for 
attacking the corona: first, visual observations by the naked 
eye, supplemented by the telescope; second, photography; 
third, polariscopic observations; and fourth, the spectro- 
scope. The polariscope had already shown that part, at 
least, of the coronal light was reflected sunlight. This con- 
clusion was corroborated by the discovery by Janssen at the 
eclipse of 1871 of dark Fraunhofer lines in the coronal 
spectrum, chief among which the D line of sodium was 
recognized. The green emission line discovered at the 
eclipse of 1869 was observed again in 1870 and 1871, 
Tennant in the latter year discovering that this ray was 
quite as conspicuous in a rift in the coronal light as in 
the adjacent streamers. The corona appeared thus to 
be shining from the luminosity of some unknown gas, which 







Photograph from Mt. Wilson Observatory. 


strange to say shone as strongly in the dark regions of the 
corona as in the bright streamers. 

In 1871 for the first time, and due to a suggestion by 
Young, a slitless spectroscope was tried. With the use of 
such an instrument at mid-totality, the emission lines ap- 
peared as rings of light, from the extent of which one could 
ascertain the height of the various gases forming the corona. 
By the help of photography in 1871, Lockyer showed that 
hydrogen extended uniformly about the sun to the enormous 
height of 200,000 miles. Could the solar atmosphere pos- 
sibly extend to this colossal distance? Let us reason by 
analogy with what we know of the earth. Its atmosphere, as 
known from observations of meteors, extends about one 
hundred miles above sea-level. The sun is more than one 
hundred times the diameter of the earth and with the same 
relative distribution the solar atmosphere might conceivably 
extend 10,000 miles from the solar surface. But an atmos- 
phere possesses mass and is obedient to the law of gravita- 
tion, and since gravity on the sun is twenty-seven times that 
on the earth, we would expect that a true atmosphere on the 
sun could not extend to elevations of more than 4,000 miles. 
But the hydrogen observed was at distances of 200,000 
miles, with the green ring reaching the still greater extent 
of 300,000 miles, whereas the coronal streamers were seen 
to stretch out several millions of miles from the sun's surface. 
Apparently, the spectroscope had not solved many of the 
difficulties of the coronal structure, but rather had succeeded 
only in complicating matters, for to add to the difficulties 
already great, it was now necessary to explain how it was 
possible that the luminous gases hydrogen and coronium 
could extend to the very great distances revealed by the spec- 
troscope. No terrestrial origin could be found for the green 
coronal line, nor did Young's discovery in 1876, that Kirch- 
hoff's " 1474 " was a double line help solve the problem. 
The perplexities were indeed very great. A faint ray of 
hope appeared in an unexpected quarter. In 1866, shortly 
after the great November meteor shower, Schiaparelli 
proved that the Perseid meteors moved in the same path as 
Tuttle's comet of 1862, while the Leonids and the Temple 


comet had identical orbits. This double coincidence between 
meteor and cometary orbits was corroborated in 1872 when 
it was found that the Andromedes, or Bielids as they are now 
called, had the same path about the sun as the lost Biela 
comet. The importance of meteors in any cosmical proc- 
ess was thus realized, and it was but natural that attempts 
should be made to solve the coronal puzzle by means of the 
meteoric hypothesis, but as we shall see later, with little 
success. Newton and Cleveland Abbe in America and 
Lockyer in England pinned most faith to the meteoric ex- 



total eclipse of July 28, 1878, was observed in 
the United States from Wyoming to Texas. On 
account of the great interest in and the vast im- 
portance of the problems to be solved, the eclipse was widely 
observed from some twelve stations by about one hundred 

Ever since the discovery of the sun-spot period in 1843 
and the finding that the earth's magnetism possessed the 
same period, astronomers had been on the alert to ascertain 
whether other solar phenomena moved in cycles parallel to 
the sun-spot curve. It had already been found that both 
prominences and faculae were more numerous when spots 
were great in number, and naturally the question was asked 
whether the coronal streamers did not originate with greater 
energy when spots were at a maximum. Here was an oppor- 
tunity to test any conclusions, for Wolf's sun-spot number 
for July 1878 was o.i, representing a minimum of spots, 
while the number for December 1870 was 135.4, a time of 
maximum of spots. 

In the state of Colorado the moon's shadow path crossed 
the Rocky Mountains. Pike's Peak (altitude 14,400 feet) 
was occupied by Professor Langley and party. At their 
station before the day of the eclipse there was hail, rain, 
sleet, snow, fog and every form of bad weather which con- 
tinued for a week, and to add to the discomforts, the horrors 
of mountain sickness had to be overcome. But on July 28, 
the weather conditions along the eclipse track were of the 
very best. 

The first change noted in the corona of 1878 was the 
enormous decrease in total lustre, when compared with the 
coronas of 1870 and 1871, Harkness estimating the 


luminosity to be only one-seventh of the corona of 1870, 
while Lockyer regarded 1878 to be one- tenth of the bright- 
ness of the 1871 eclipse. The decrease in brilliancy was 
accompanied by a remarkable and unexpected change in 
shape. To Janssen in 1871, the dark moon looked like the 
center of a giant dahlia, the corona being nearly circular in 
outline. In 1878, the streamers along the sun's axis were 
much shorter in length but much more pronounced in char- 
acter, these polar rays resembling more than anything else 
the lines of force around a magnet. But the most astounding 
phenomenon was the enormous extent of the coronal stream- 
ers along the sun's equator. Langley in the pure, rare air 
of Pike's Peak followed these streamers to six diameters of 
the moon on one side, but on the other side where he had 
been more intently watching, to the colossal length of twelve 
diameters, or more than ten millions of miles! These 
equatorial extensions were confirmed by Simon Newcomb in 
Separation, Wyoming, by Cleveland Abbe farther down the 
slope of Pike's Peak, and by almost every astronomer who 
witnessed the eclipse. The perplexities surrounding the 
corona were accordingly multiplied many-fold, for how- 
could a solar atmosphere obeying gravity exist at the huge 
distance of ten million miles from the sun's surface? Young 
and Abbe saw long faint beams shining along the sun's 

Remarkable as were the visual phenomena manifested, 
their testimony was no whit stranger than the revelations 
by means of the spectroscope. The hydrogen and green 
coronium emission lines were visible, but with such vastly 
diminished intensity compared with the eclipse of 1871 that 
most observers completely missed seeing them. They were, 
however, visible to Young, Eastman and some others. If 
the emission lines were weak, the Fraunhofer lines of the 
corona were comparatively strong, showing that the re- 
flected light near the sun's limb was relatively stronger than 
in 1871, a fact confirmed by observations with the polari- 

The eclipse of 1878 showed the long equatorial wings 
of the corona, strong polar brushes, faint incandescent light 


of coronium and hydrogen, and light reflected strongly 
from material particles near the sun's limb. Were each of 
these four special features unalterably connected with the 
condition of the sun-spot minimum, or did they happen 
merely by chance? Time alone could furnish the answer. 
Great progress was made at this eclipse in the photography 
of the corona, particularly by the use of portrait lenses which 
were successful in portraying a mass of detail in the inner 
and brighter corona, but failed to show the outer streamers. 
Photographs of these faint extensions must needs wait 
until some date in the future when plates of greater sensi- 
tivity could be produced. 

Another important observation at the eclipse of 1878 was 
the discovery (?) of two bright star-like objects by two 
American astronomers, Swift and Watson. The objects 
could not be identified with any of the fixed stars, and it 
was therefore necessarily assumed that they were small 
planets moving about the sun inside of the orbit of Mer- 
cury. The reputations of these two astronomers for care- 
ful observing were so great that it cost the science of as- 
tronomy a quarter of a century of eclipse observations 
before it was finally decided that no intra-Mercurial planets 
exist which are as large or as bright as the objects supposed 
to have been seen. 

The next eclipse to be observed was that of May 17, 
1882, the forerunner in the Saros of the eclipse of May 28, 
1900. The 1882 eclipse was seen in Egypt with a brief 
duration of totality amounting to seventy-four seconds. 
This eclipse is memorable on account of the bright comet 
that was seen and photographed near the sun, the comet 
not being observed either before or after the eclipse. The 
photographic plates had now become more rapid, the dry 
plate having been invented, and accordingly the astrono- 
mers had to their hands better facilities for attacking the 
corona with camera and spectroscope. Also for the first 
time we hear of the prismatic camera, which is a slitless 
spectroscope, with a photographic plate to take the place 
of the observing eye-piece; and this instrument, particu- 
larly in the hands of Lockyer, was to play an important role 


in eclipse spectroscopy. Eleven years having elapsed since 
1871, the year 1882 was one of maximum sun-spots, there 
being no less than twenty-three separate spots on the face 
of the sun on the day before the eclipse. The form of the 
corona in no way resembled that of the minimum of 1878 
but bore a striking resemblance to the crown of glory of 
1871, the shape being more nearly rectangular , or even star- 
like, and the long equatorial extensions and strong polar 
brushes being entirely lacking. The spectroscope also re- 
vealed vast differences from the eclipse of 1878. The corona 
as a whole was more brilliant than that of the preceding 
eclipse, and the emission lines of the spectrum were obtained 
both by a slit spectroscope and by the prismatic camera. 
With the former instrument, Schuster photographed about 
thirty lines in the spectrum of the corona. Many new spec- 
tral lines were visible in the red and violet to Tacchini and 
Thollon respectively. These lines were seen and photo- 
graphed during the progress of totality, and not near the 
beginning or end of the total phase. Apparently the lines 
did not seem to belong to the flash spectrum and must have 
their origin in the true corona. But for the first time a sus- 
picion seems to have been aroused that the lines might after 
all be due to prominences and chromosphere and not to the 
true corona, for Schuster observed the H and K lines of cal- 
cium to appear bright even across the face of the dark moon 
where no light at all was supposed to exist! Evidently the 
chromospheric light was reflected by some atmosphere 
somewhere, either in the higher reaches of the sun's atmos- 
phere directly in line with the center of the dark moon, 
or in the atmosphere of the earth. The first condition could 
hardly be possible and that left no contingency other 
than the second. Schuster's observation was not the first 
to reveal bright lines on the dark face of the moon because 
as early as 1870 Young had perceived bright hydrogen lines. 
If the lines of emission were stronger in 1882 than in 
1878, it was not so with the dark Fraunhofer lines. The 
spectroscopes revealed the continuous spectrum in the 
brighter inner corona, but farther from the sun the dark lines 
due to reflected photospheric light were observed both visu- 


ally and in the photographs, but these lines were not so strong 
as in 1878. Sun-spot maximum appeared therefore to 
correspond to a star-like corona, with no polar brushes, 
strong coronium and other bright coronal lines, but with the 
Fraunhofer lines intrinsically weaker than at sun-spot 

The direct photographs of the corona by Schuster were 
in better definition and showed more details than those of 
previous eclipses. In fact, the impressions made on the 
plates were so strong that Dr. Huggins obtained the idea 
that it might even be possible to photograph the corona 
without an eclipse. For observing the prominences without 
waiting for an eclipse, the spectroscope had already been 
utilized to get rid of the glare of the sunlight in our own 
atmosphere. It is not possible to make use of the spectro- 
scope in the same manner for obtaining coronal photographs, 
for the simple reason that the bright-line radiations of coro- 
nium are not sufficiently strong in character to enable the 
coronium light to shine in contrast with the enfeebled solar 
glare. Many attempts to photograph the corona in full sun- 
light were made by a variety of different methods lasting 
over a number of years, the astronomers being urged on by 
great hopes since the first trial photographs seemed to pre- 
dict success. The details of the early work of Huggins, and 
the later researches will be deferred until a subsequent 

One-third of the way round the globe eastwards from the 
Dutch East Indies, where the eclipse of 1901, May 18, was 
observed, brings one to the middle of the Pacific Ocean. 
The eclipse track of May 6, 1883, lay almost entirely across 
the water, but fortunately, in its path there was a small coral 
reef, only seven miles long, and unknown ten years previ- 
ously. The importance of the discoveries of 1882 and the 
fact that the eclipse of the year 1883 was of the very long 
duration of more than five minutes, attracted to Caroline 
Island, astronomers from America, England, France, Austria 
and Italy. The great risks taken by eclipse expeditions in 
the tropics of being overtaken by clouds was shown in 
this eclipse, but fortunately a clear spell between two 


periods of clouds was experienced. The general features 
of the corona greatly resembled that of the year before, the 
sun continuing to show many spots. Owing to the long ex- 
posures possible, excellent photographs of the corona were 
secured, and for the first time in the history of eclipses, 
greater extensions of streamers were photographed than 
were visible to the eye. 

The most important observations were unquestionably by 
means of the spectroscope. Up to this time, all observa- 
tions during totality had shown a continuous spectrum of 
the corona close to the sun, and farther out faint Fraun- 
hofer lines, with the bright green line of coronium crown- 
ing the whole. According to the dissociation theory of 
Lockyer, neither continuous spectrum nor dark lines could 
exist there, and " if these statements regarding the corona 
were strictly accurate my hypothesis was worthless/' 1 
Hence, a careful search was made by Janssen for Fraun- 
hofer lines. They were found by him in great numbers, 
thus confirming the observation of 1882. To make as- 
surance doubly sure, the dark spectrum lines in the corona 
were successfully photographed. As a result of these ob- 
servations, Janssen concluded 2 that " the basis of the coro- 
nal spectrum was formed by the complete Fraunhofer spec- 
trum, and that, therefore, there exists in the corona, and 
above all in certain localities of it, an enormous amount of 
reflected light; and since we know that the coronal atmos- 
phere is very rare, it follows that these regions must 
abound in cosmic matter in the state of solid corpuscles, in 
order to explain the abundance of reflected sunlight." 

Spectroscopic observations of great interest were made 
on the corona by Hastings. He used a 60 prism attached 
to a six-inch telescope, there being two totally reflecting 
prisms placed outside the slit so that the spectrum of two 
opposite sides of the sun could be brought together and ex- 
amined by comparison. The observations were confined to 
the green coronium line. At the beginning of totality, this 
line was 12' in length and very bright on the eastern 

1 Lockyer, Chemistry of the Sun, p. .365. 

2 Comptes Rendus, 92, No. 10. 


limb, while on the western limb it was only 4' in length and 
comparatively faint. As the eclipse advanced , the inequal- 
ity vanished, at mid-totality conditions were equal, while 
at the end of totality the lines on the western limb were the 
longer and brighter. Such a great change could not be ex- 
plained by assuming that the moon in its motion progres- 
sively covered and uncovered the bright coronal radiations, 
and accordingly, Hastings attempted to explain his observa- 
tions on the assumption that the outer corona has no real 
existence but that its appearance is caused by diffraction 
round the edge of the moon. On this hypothesis, the true 
corona is confined to a very narrow ring around the sun, 
the light from this inner ring of material substance being 
widened by diffraction to form the outer corona which thus 
takes upon itself all of the appearances of reality. To the 
astronomers who had seen the great extensions of 1878, it 
was hard to believe that diffraction of light could adequately 
explain the detail of the coronal streamers at the great dis- 
tance of twelve diameters from the sun's limb, but it was 
equally difficult to understand how luminescence could 
exist in a solar atmosphere at the colossal distance of ten 
million miles from the sun's surface. If the coronal light 
were reflected, it could not be seen unless reflected from 
material particles, and if the light were intrinsic, how could 
it have any existence in an atmosphere so infinitesimally 
rare? The answer given by Hastings denied the solar origin 
of the corona, and seemed to be a step backward. Appar- 
ently there was no way out of the quandary, but to wait for 
future eclipses. 

In the attempt to secure information regarding the flash 
spectrum three distinct improvements in the line of attack 
were inaugurated in 1882: i. Eye observations were no 
longer to be trusted exclusively, the spectrum must be 
photographed. 2. A grating was used, before, during and 
after totality. 3. A moving plate was utilized with an 
integrating prism spectroscope to secure photographs before, 
during and after totality. The grating showed little in ad- 
dition to the H and K lines seen near the limb throughout 
totality. The photographs attempting to secure the revers- 


ing layer succeeded only in imprinting the hydrogen lines 
and comparatively few of the brighter lines, in fact the 
total number of lines were certainly too deficient to guar- 
antee the confirmation of Young's belief that the whole 
Fraunhofer spectrum was reversed in the flash spectrum. 

The total eclipse of August 29, 1886, was visible in the 
West Indies. The energies of many of the observers were 
devoted to testing the method of Huggins of photographing 
the corona without an eclipse. For this purpose, fifteen 
separate photographs were taken on the day before the 
eclipse and a series of twenty during the partial phases, 
these photographs to be compared with plates obtained 
during totality. The conclusions were quite definite, for 
not a single one of the coronal details was found on the 
plates taken outside of totality, and it seemed therefore nec- 
essary to decide that it was impossible to photograph the 
corona, except within the limits of a total eclipse, at least 
under the conditions of hazy sky and low sun that had pre- 

Tacchini made a careful comparison of the prominences 
observed spectroscopically before and after totality with 
those seen directly during the total phase, and he concluded 
that all the prominences showed themselves larger and 
taller during an eclipse, the upper portions being white in 
color when the prominences exceeded i' of arc in height. 
The differences of apparent height may find a ready ex- 
planation in the effect of contrast with the background, in- 
side and outside of totality, but the matter of the color of 
the prominences could not be so readily settled. For many 
years " white " prominences found a conspicuous place in 
spectroscopic literature. Tacchini observed the flash 
spectrum visually. Turner attempted to observe changes 
in the coronal streamers resembling currents, but obtained 
no results of value. 

One of the expeditions at this eclipse experienced a series 
of accidents due to the unavoidable necessity of employing 
volunteer observers as assistants. Some of the incidents 
might have been amusing if an opportunity had been af- 
forded on the morrow for making another attempt, but 


at the rare event of a total eclipse such untoward happenings 
are not laughing matters, but become almost tragedies. 
Some of these mishaps were: failure to get the sun's image 
on the photographic plate in the most important instrument, 
the breaking of the polar axis just before totality, in the 
next important instrument, the failure of an assistant to 
make exposures, the standing of two native policemen in 
front of the photometer during totality, the seizure of the 
plates by well-meaning but ill-advised customs officials, 
thereby causing a delay in the development of the plates 
until the following May by which time the plates had 
greatly deteriorated. Eclipse observers generally are not 
anxious to repeat such a chapter of accidents. 

The eclipse of the following year, August 19, 1887, was 
one of widespread disappointment, for the projects so care- 
fully prepared ended only in failure to secure results, not 
through any fault in the plans themselves but on account of 
the astronomer's enemy, clouds, which prevailed almost 
everywhere. Fine weather, due to holes in the cloudy sky, 
prevailed at several of the stations in Russia and Japan, 
however, and some photographs and observations of value 
were secured. 

The year 1889 brought two eclipses, both extensively ob- 
served. Here was inaugurated the splendid series of ex- 
peditions sent out from the Lick Observatory. The path of 
the eclipse of January i crossed Nevada and California, and 
the photographers near the line of totality were so well or- 
ganized for the work by Mr. Charles Burckhalter that an 
excellent series of photographs of the corona resulted. 
The best photograph secured at this eclipse, and in fact the 
very best obtained at any eclipse to this date, was secured 
by Barnard. The equipment was very meager. The largest 
lens employed was 3^ inches aperture, stopped down to i^ 
inches, and of 49 inches focus. Barnard's success depended 
on an accurate adjustment of the instrument but more 
specially on the skill and care with which the plates were 
developed. Barnard was a professional photographer be- 
fore he was an astronomer, and he was thoroughly familiar 
with the best methods of developing a plate in order to 


bring forth all of the latent detail. The eclipse of Decem- 
ber 22 of the same year was successfully photographed at 
Cayenne in the West Indies by the Lick party consisting 
of Burnham and Schaeberle. The photographs showed that 
changes had occurred in the corona since the eclipse of the 
beginning of the year. The earlier eclipse took place near 
sun-spot minimum and exhibited the equatorial extensions 
of the corona. The eclipse of December 22 is memorable 
from the death of Father Perry a few days after the eclipse, 
a martyr to the cause of science. This brave man, though 
greatly weakened, took part in the eclipse work, and having 
found as soon as totality passed that everything had passed 
off well, he called for three hearty British cheers, which 
unfortunately he could not himself lead. 

The greatest success attended the observations of April 
1 6, 1893, largely through the use of apparatus much more 
powerful than had ever been employed before at an eclipse. 
The most conspicuous advance came to the party from the 
Lick Observatory in Chile who used a camera of five inches 
aperture and forty feet focus for securing photographs of 
the corona. Schaeberle decided to point the objective di- 
rectly at the sun and to mount it on one fixed pier and the 
movable photographic plate on another, both piers to be 
wholly free from contact with the great tube extending 
between lens and plate. The slide carrying the photo- 
graphic plate was the only moving part, and its motion was 
so regulated by means of inclined planes as to give it the 
same velocity and direction as the sun's focal image during 
the eclipse. The details of erecting this instrument, known 
as the Schaeberle mounting, are found in the Contributions 
from the Lick Observatory, No. 4. A careful focus was 
secured and beautiful photographs were obtained showing 
the prominences and inner corona with a definition which 
left little to be desired. 

Optical power, up to then unprecedented, was employed in 
the prismatic camera designed by Lockyer and used by 
Fowler in West Africa. The camera had a focal length of 
7 feet 6 inches with a prism of 45, giving a dispersion of 
about two inches from F to K. Fowler secured photographs 


which were supplemented by a series obtained by Shackleton 
in Brazil, with the result that the positions of 164 chromo- 
spheric lines were measured between F and K. Deslandres, 
at the same eclipse, attempted to measure the rate of rotation 
of the corona by observing the relative displacement of the 
spectra of two regions of the corona at opposite sides of 
the sun placed in juxtaposition. A grating spectroscope 
was used, and the conclusion reached was that the corona 
partakes of the general rotation of the sun. Unfortunately, 
there is no justification for this deduction by Deslandres 
since the measures made by him were on the H and K lines 
which belong to the chromosphere and not to the corona. 
One of the most important results of this eclipse was that 
it became possible for the first time in eclipse spectroscopy 
to separate clearly the spectrum of the corona from that 
of the chromosphere, and it was henceforth no longer as- 
sumed that a spectral line visible during totality belonged 
of necessity to the corona. 

The eclipse of 1896, taking place on August 9, was ob- 
served by a large number of expeditions. An English party 
consisting of Christie, Turner and Hills went to Japan 
where also was one from the Lick Observatory and another 
American expedition headed by Todd ; and also two Japa- 
nese parties. 

Lockyer went in H. M. S. Volage to Norway where 
a large party of seventy-five, including officers and sailors 
of the ship, took care of a large and varied program. In 
one department of the work, for instance, in the sketching 
of the corona, a competition was started on board ship by 
thirty-five volunteer sketchers, an artificial corona being ex- 
posed to view for 105 seconds, the time of duration of 
totality. Sixteen who showed the greatest proficiency were 
selected for sketching of the corona on eclipse day. But 
alas, for " the best laid schemes o' mice and men," the 
clouds prevailed almost everywhere except where there was 
a small English party consisting of Stone and Shackleton. 
The latter was successful in timing his observations with 
the prismatic camera so well that the long desired photo- 
graph of the flash spectrum was at last secured. Although 


the focus was not of the very finest, still there were shown a 
total of 464 lines in the spectrum between F and K. This 
photograph, taken by one of Lockyer's assistants, sounded 
the death-knell of the dissociation theory, but Lockyer 
still refused to be convinced that the flash spectrum was a 
reversal of the Fraunhofer spectrum. His argument was a 
very simple one, which was, that between F and K, 5694 
Fraunhofer lines were tabulated by Rowland, while in the 
eclipse spectra of 1893 there were but 164 lines and in 1896 
but 464 lines, consequently showing but three and eight 
percent, respectively, of the Fraunhofer lines reversed in 
the flash spectrum. Lockyer however failed to draw atten- 
tion to the fact that Rowland's atlas was secured with a 
much greater dispersion than that used at the eclipse and 
with vastly superior definition. Lockyer's conclusion, 1 the 
result of the spectra at these two eclipses, was that " the 
chromosphere is a region of high temperature in which there 
is a corresponding simplification of spectrum as compared 
with the cooler region in which the Fraunhofer absorption 
is produced." The manner of settling the question, the way 
of advancement for future eclipses was clearly indicated: 
the flash spectrum must again be photographed and with 
increased dispersion, and great care should be exercised to 
see that the exposures were made at the correct times, with 
as good focus and definition as possible. 

Such photographs were secured at the eclipse of January 
22, 1898, visible in India, where such excellent conditions of 
weather were experienced that then partially compensated 
for the ill luck of the previous eclipse. The largest expedi- 
tion in point of numbers was that under the direction of Sir 
Norman Lockyer located at Viziadrug on the West Coast, 
the astronomers being assisted by the officers and men of 
H. M. S. Melpomene. The program was an extensive one, 
embracing visual and photographic observations of the 
corona, the most important problem being a spectroscopic 
attack on the chromosphere with two large prismatic cameras 
of six and nine inches aperture. By these two instruments 
about sixty photographs were secured, the exposure times 

1 Recent and Coming Eclipses, p. in. 


varying from i to 59 seconds. These included two series of 
ten snap-shots at the beginning and another ten at the end of 
totality and a number of exposures of different lengths dur- 
ing totality. Christie and Turner, representing the British 
Joint Permanent Eclipse Committee, were at Sahdol. Cope- 
land was at Goglee, Newall and Hills were at Pulgaon, 
Campbell of the Lick-Crocker expedition was at Jeur, while 
at Talni was located Evershed and also Mr. and Mrs. 

The most important problem was that of the flash spec- 
trum, and fortunately, successful photographs were secured 
by Fowler and Dr. Lockyer, by Campbell, by Hills, by 
Newall and by Evershed.. A discussion of the spectra by Sir 
N. Lockyer again confirmed him in the opinion he had held 
since 1873, that many strong chromospheric lines were not 
represented among the Fraunhofer lines, while many of the 
dark . lines found under ordinary conditions in the solar 
spectrum did not appear as bright lines in the flash spectrum. 
He therefore concluded that the flash did not represent 
the spectrum of the reversing layer. It is true that the hy- 
drogen series and the helium lines of the flash spec- 
trum are not found in the Fraunhofer spectrum, and also 
that there are great differences in intensity between the two 
spectra, and in a sense therefore one spectrum is not the 
exact reversal of the other; but none the less, it is impossible 
to reach any conclusion other than that practically every 
strong dark line in the solar spectrum is present as a bright 
line in the flash spectrum. The matter will be treated more 
fully in Chapter XVI. The excellent photographs obtained 
by Fowler and Lockyer permitted a discovery of the greatest 
value to future scientific development, namely, that " en- 
hanced " lines, or those which are relatively stronger in the 
spectrum of the spark than in the arc, are specially promi- 
nent in the spectrum of the chromosphere. 

Exquisite photographs of the corona with the 4O-foot 
camera were secured by Campbell, while Mrs. Maunder 
with a Dallmeyer lens of only one and a half inches aper- 
ture photographed the faint extensions of the corona run- 
ning out to nearly six diameters from the moon's limb. The 


corona of 1898 presented a mixed aspect, a combination 
of the polar brushes observable at sun-spot minimum being 
combined with the quadrilateral shape of sun-spot maxi- 

One contribution of great importance was the measure- 
ment of the wave-length of the green coronium line. For 
nearly thirty years since its discovery, it had been assumed 
that the coronium line was identical in position with the 
chromospheric line at 5316.8 A. Lockyer, Fowler, Evershed 
and Campbell independently found the coronium line to be 
farther to the violet, at 5303 A. That the value of this 
wave-length was now found for the first time, in spite of 
observations made at several eclipses, will show more clearly 
than words can express that the eclipse spectra prior to 1898 
were poor in definition and small in dispersion. 

Newall used a spectrograph with two slits with which he 
hoped to secure photographs to test the rotation of the 
corona. Unfortunately, the slits were placed 8' from the 
sun's limb and the coronal light was too feeble to impress 
any traces on the plate. Newall observed the corona with 
a polariscope while Turner attempted to achieve similar ob- 
servations by photography. 

A new epoch of accuracy in photographing the chromo- 
spheric spectrum having been begun in 1898, it was but 
natural that every effort should be made to continue the 
success of this work in 1900, in order to secure, if possible, 
still greater definition with larger dispersion. The eclipse 
track of May 28, 1900, lay over the southeastern part of 
the United States, and after crossing the Atlantic Ocean, 
passed over Portugal, Spain and Algeria. On account of its 
easy accessibility to American and European astronomers 
the eclipse was witnessed by a greater number of observers 
than ever before in the history of eclipses. Fortunately 
good weather was experienced almost everywhere. The 
program was a wide and varied one, and it will be possible 
here to mention only a few special lines of work and record 
the names of comparatively few of the many observers. 

Photographs of the corona were taken in numbers to the 
hundreds, or even to the thousands, by small, medium, 



Cfl .3 -M 



S S 

3 o> 
y a 
o a 




large and huge sized cameras, the greatest focal length be^ 
ing that employed by the party from the Smithsonian In- 
stitution who utilized a lens of twelve inches aperture and 
135 feet focal length. The photographs on this large scale 
were in good definition and they showed a great wealth of 
detail in the inner corona. Excellent photographs were se- 
cured at Wadesboro, N. C., by Professor Barnard and Mr. 
Ritchey with a horizontal camera of 61^ feet focus. A re- 
production of the exposure of thirty seconds is given on 
page 1 64 . This eclipse having taken place near spot-minimum 
shows strong polar brushes and long equatorial streamers. 

Both in America and in Europe, an extended attack was 
made on the flash spectrum by slit spectroscopes, by pris- 
matic cameras and by gratings, plane and concave, used both 
with and without a slit. The United States Naval Observa- 
tory had three stations in the field, at Barnesville, Ga., at 
Griffin, Ga., and at Pinehurst, N. C. The three concave 
gratings used by Ames, Crew, and Humphreys were each 
employed with a slit. The photographs showed nothing. 
With plane grating and quartz lens, Huff secured a well-ex- 
posed spectrum, but the focus was not of the very best. 
Better results were secured by Frost who used a concave 
grating, objectively without slit. 

In Europe, Sir Norman Lockyer, located at Santa Pola 
in Spain, carried out an extended series of observations 
similar in scope to those made at Viziadrug in 1898. The 
nine-inch prism was combined with a camera 20 feet in 
focal length in order to secure a great linear dispersion in 
the resulting photographs for the purpose of measuring the 
heights of the various layers forming the sun's chromo- 
sphere. Successful photographs of the flash spectrum were 
secured by Fowler and Dr. Lockyer, and also by Dyson at 
Ovar, and Evershed at Mazapan in Algeria. The latter 
station was selected so that it might be as near as pos- 
sible to the edge of the band of totality so that the photo- 
graphs of the chromosphere might be obtained in high 
solar latitudes. Unfortunately, through an error in the 
Nautical Almanac, Evershed found himself just outside, in- 
stead of barely inside, the path of totality. The series of 


photographs obtained, however, were of fine definition and 
were specially valuable in affording a means of comparison 
with photographs of the flash spectrum which have usually 
been taken near the solar equator. This comparison shows 
that the spectrum of the chromosphere is the same at the 
sun's polar regions as at low latitudes, and it appears fairly 
certain that the spectrum of the sun's limb is as constant 
in character as the ordinary Fraunhofer spectrum. In this 
connection it should be borne in mind that Evershed's photo- 
graphs showed the flash spectrum where the moon was prac- 
tically at grazing incidence with the sun, and consequently 
the layer of the chromosphere photographed must have been 
very close to the edge of the photosphere. 

Fortunately, the Algiers Observatory was in the shadow 
track, and here through the kindness of the director, M. 
Trepeid, Mr. Wesley was given an opportunity to observe 
the corona visually with an equatorial coude of 0.3 meters 
aperture. Mr. Wesley's long experience in the study of the 
corona had well fitted him for the task of finding out whether 
more detail could be observed visually than is portrayed on 
the best photographs. The conclusion reached was very 
definite that the photograph, if on a sufficiently large scale, 
exhibits all of the coronal features that can be seen with 
the eye. 

Successful polariscopic observations were made by 
Turner, Newall, and others. The inner corona showed 
marked polarization, a result which was difficult to reconcile 
with the absence of Fraunhofer lines in the spectrum of the 

Interesting observations were made by Abbot with the 
bolometer in measuring the radiation of the corona at dif- 
ferent distances from the sun's limb. Results of great value 
were secured, but they showed the necessity of confirming 
them by observations at succeeding eclipses. 



first total eclipse witnessed by the author was 
I that of May 28, 1900. The moon's shadow path 

JL in America traversed the southeastern states from 
New Orleans to Norfolk; it then crossed the Atlantic Ocean, 
passed over Spain and the Mediterranean and left the earth's 
surface at sunset in Northern Africa. Good weather con- 
ditions had been predicted on both sides of the Atlantic and 
so it was unnecessary for the European astronomers to 
travel to America or for the Americans to cross the Atlantic 
eastward in order to insure a promise of successful opera- 
tions. A few astronomers, however, did make the sea 
journey. As an illustration of how little the average 
European knows about America or rather it should be 
said, did know of the United States in the year 1900 it 
can be said on good authority that before crossing the 
Atlantic to witness the eclipse, the head of one of the ex- 
peditions appealed to the United States government for a 
guard of soldiers to protect the lives of the party from the 
wild natives (sic) of North Carolina. On arriving in New 
York, however, the fears were effectually dispelled when 
the party found themselves aboard a luxurious Pennsylvania 
railroad train and discovered that they themselves, their 
baggage and their instruments were routed through to their 
destination, and entirely free of charge. 

In view of the accessibility of the eclipse track in the 
United States and on account of the fact that it was a very 
favorable season of the year for a trip to the South, it was 
but natural that such great American institutions as the 
Lick, Yerkes and Naval Observatories should send well- 
organized expeditions, and that individual astronomers 
should gather in great numbers to witness the fascinating 



phenomenon. Most of the American astronomers were 
seeing their first eclipse and it might well have been thought 
that they would suffer from " nerves " at the exciting 
moments of totality; but there were many seasoned veterans 
in the ranks, men like Young, Langley and Campbell, who 
had witnessed eclipses in foreign lands. 

Perhaps the most important part of the program for 1900 
was the spectroscopic work. In 1896, after many trials, the 
first good photograph of the flash spectrum was obtained by 
Shackleton. In India in 1898, as a result of better defini- 
tion, the spectrum photographs exhibited a wealth of detail 
that added greatly to our knowledge of solar physics. Evi- 
dently the procedure for the next eclipse should be the 
attempt to secure spectra with increased dispersion, so that 
wave-lengths should be determined with greater accuracy, 
thus permitting more reliable information regarding the 
sources of the spectral lines and rendering more positive 
our knowledge concerning motions of rotation of the solar 
envelopes. America is the home of the diffraction grating, 
plane and concave, brought to such perfection by the refined 
labors of Professor Henry A. Rowland of Johns Hopkins. 
Among the more powerful instruments brought into service 
were three concave gratings, each used with slit, and two 
plane gratings each used objectively without slits. Each of 
these five gratings were employed in connection with a 
quartz lens. Ordinary photographic plates were used, and 
as these are but little sensitive to green, yellow and red 
light, it was decided to concentrate on the blue and violet 
parts of the spectrum; hence the employment of quartz, 
rather than glass, lenses. The concave gratings were 
mounted in the ordinary Rowland manner in the attempt 
to secure sharp definition and large dispersion but the 
dispersion was too great for the light available and no lines 
were found on the photographs. Huff obtained well-exposed 
plates with a plane grating, but the focus was good for a 
short region only near wave-length 4000 A. 

The experience gained in 1900 was put to use in the plans 
for the eclipse of the following year. On May 18, 1901, the 
moon's shadow touched the earth's surface at sunrise on the 






east coast of South Africa. After passing over the Island of 
Mauritius, the shadow traversed the Indian Ocean at cannon- 
ball speed and touched land shortly after noon on the west 
coast of the Island of Sumatra in the Dutch East Indies. 
After crossing over northern Borneo, Celebes and New 
Guinea, the shadow left the earth's surface at sunset far 
south of the Philippines. This eclipse was specially im- 
portant on account of the very long duration of totality, 
six minutes. The location that afforded the best living con- 
ditions and promised the greatest success from favorable 
weather was the west coast of Sumatra, but such a site 
would carry the American astronomers as far away from 
home as they could possibly get half way round the world. 

In view of this great duration of totality and the im- 
portance of the observations to be attempted, the United 
States government, through Congress, appropriated money 
to equip and send out an expedition of thirteen members, 
two from the Smithsonian Institution, six from the U. S. 
Naval Observatory, and five guests of the latter, all of whom 
had had eclipse experience in 1900, the author being one of 
the five guests. In order to arrive in the East Indies in 
plenty of time before the day of the eclipse, May 18, the 
expedition left San Francisco on February 16 on board the 
U. S. Army Transport Sheridan, en route to Manila. On 
the journey westward three delightful days were spent in 
Honolulu, viewing the points of interest of that far-off 
" Paradise of the Pacific. " At that time the Hawaiian 
Islands were cut off from communication with the outside 
world, except in so far as news was brought by steamer, for 
there was no cable to Honolulu and it was before the 
days of wireless. 

On March 26th, the U. S. S. General Alava left Manila 
to carry the expedition the remaining 2200 miles to the 
west coast of Sumatra where it was intended to locate for 
the purpose of the eclipse observations. After coasting 
along the Philippines and the north shore of Borneo, the 
equator was crossed on March 3ist. Father Neptune came 
on board, and the reception to him was right royally given 
by the man-o'-war's men. April 2nd found us in the straits 


of Sunda which separate Java from Sumatra, and the ship 
passed within half a mile of the island volcano Krakatoa 
whose eruption in 1883 resulted in the loss of 40,000 lives. 
The huge tidal wave following the explosion, which inundated 
the whole surrounding country, left a Dutch man-of-war high 
and dry a mile and a quarter inland, and 76 feet above high 
water mark ; and this wave was felt even in the English Chan- 
nel, 11,000 miles away. The air wave sent out by the erup- 
tion was traced by barometers through seven complete cir- 
cuits of the earth. Fine particles of dust were shot up to 
enormous distances in the atmosphere, causing the brilliant 
sunsets noticed for many months in 1883. 

After coasting along the west of Sumatra, we entered the 
pretty little land-locked harbor of Emma Haven, probably 
the first American government ship that had ever entered 
port there, and were then at the end of our sea voyage of 
over 1 0,000 miles. 

After official calls on the governor of that peaceful Dutch 
colony, we proceeded to Padang, the capital, just four miles 
distant, and obtained our first view of the island which was 
to be our home for the next two months. First impressions 
were formed at the Oranje Hotel, a typical hostelry of the 
East. The building, half hidden by cocoanut and tropical 
palms, is of one story with high thatched roof. At the front is 
the wide open verandah the parlor and reception room 
of the hotel and back of this, running through the middle 
of the building, is the dining room. On either side are the 
spacious bed rooms, each with its wide verandah exceed- 
ingly inviting places in the hot afternoons. The bed is the 
chief curiosity to the American, for it is a four-poster, with 
finely-figured mosquito net and of the huge dimensions of 
7x8 feet. Over the mattress is spread a daintily embroid- 
ered sheet, but there are no bed-covers and indeed, under 
the equator, few blankets are needed. Of pillows, there are 
several, the ordinary bed having four and two " Dutch wives " 
the name given to a long bolster used to keep the occupant 
of the bed cool. 

A day in the East is begun by awaking about six o'clock. 
Coffee is immediately brought by the jungus for no white 


man does any manual labor, and each has his native valet 
and then, clad in pajamas, one risks the bath. Such a 
thing as a bath tub is unknown, the bath being taken in a 
cemented room with a cistern of water at the back, over 
the top of which is placed a wire screen to keep out inquisi- 
tive foreigners. Dipping water through a rectangular open- 
ing in the screen by means of a bucket, and throwing it over 
oneself, constitutes the bath and a very excellent one it 
proved to be. After leisurely dressing (for no one is ever 
in a hurry), a light breakfast is partaken of, consisting usu- 
ally of bread and a few cold meats. The men come to break- 
fast dressed for the business of the morning, in their white 
suits, or often in their pajamas, consisting of trousers of 
cotton stuff of the most marvelous colors and patterns, and 
a coat or kibaya of white, made after the Chinese style 
without collar. But how tell of the dress of the ladies? For 
they appear at breakfast in costumes in which an American 
girl might be ashamed to be caught in her boudoir: a native 
skirt or sarong reaching to the ankles made of picturesque 
cotton stuff, and a lace-edged linen jacket. This is the dress 
of the native women who wear neither shoes nor stockings. 
The Dutch women have adopted their costume in toto, except 
that they usually add a pair of gold embroidered slippers. 
After breakfast, the serious business of the day is taken up 
until it is time for the midday meal, or rijsttajel. This, as the 
name signifies, is largely made up of rice and it is a most 
astonishing meal. Into a large soup plate is put a liberal 
supply of splendidly cooked rice, and on top of this, chicken, 
meat, potatoes, and portions of fifteen or twenty curried 
dishes. It makes an astounding looking mess, but after a 
little experience in selecting the proper proportions and 
combinations of the curry and spices, the whole makes a very 
palatable dish. This is only the first course, to be followed 
by potatoes fried in cocoanut oil, and excellent chicken, with 
a third course of fruit. The next two hours of the day are 
universally given up to the siesta. About four o'clock, the 
East again awakes, and after a cup of tea or chocolate, takes 
another bath and dresses for the afternoon. The Dutchman 
seeks his club and plays a game of billiards or cards. Games 


requiring much exercise do not find favor in his eyes, but he 
has a taste for horse flesh, breeds most wonderful little ponies, 
and is always riding or driving. The ladies now appear for 
the first time during the day in European costyme ; and the 
hours before late dinner are given up to social functions. If 
these are of a ceremonious nature, the men must wear a black 
tail coat, and at least carry a hat, even if it is not worn. 

A nice little informal gathering takes place before dinner 
each evening on the hotel verandah, where the men drink 
their pitje of gin and bitters, furnished by the hotel to all 
comers. Dinner at 8:30 or 9:00 has nothing remarkable 
about it, being very similar to our own. As soon as dinner 
is over, negligee is resumed, and the day is at an end. 

Within a week there arrived in Padang two other parties 
besides our own from the United States, and scientists from 
England, Holland, France, Russia, Japan, and even India, 
about eighty astronomers altogether from all parts of the 
world, all bent on learning something about the sun. 

The Dutch were very kind and generous in their treat- 
ment of the foreign scientists, and did everything in their 
power to make the two months' stay in their island as easy 
in becoming acquainted with strange conditions and as en- 
joyable, as possible. Passes were furnished to each astrono- 
mer for passage over the " Staatsspoorweg op Sumatra/' the 
railroad owned by the government which runs from Padang 
into the interior about 100 miles, one spur reaching Fort de 
Kock, the other Sawah Loento. These passes included not 
only the transportation of our persons, but also of freight 
and baggage, which is not carried free in the island. Labor 
was largely furnished without cost, machine shops were put 
at the disposal of the astronomers, and their slightest wish 
was readily met and generally anticipated. In short, these 
Hollanders were perfect models of hospitality. 

In a few days after arrival, the astronomers had separated 
to the different locations determined on as bases of opera- 
tions. The English divided into two parties, one going to 
Sawah Loento, where were located Mr. and Mrs. Newall, 
the other to an island off the coast of Sumatra where the 
party under Dyson could have the assistance of the officers 


and men of H. M. S. Pigmy, detailed there to help in the 
observations. The Dutch located not far from this island 
but on the mainland, and the astronomers had the help of 
the gunboat Sumatra. Perrine of the Lick Observatory of 
California decided on Padang as likely to be most favorable 
for his researches. 

The Naval Observatory divided their party into three sec- 
tions, and decided to locate along the line of the railroad. 
The engineering principles under which this road was con- 
structed were marvels of simplicity. If a small hill was met 
in laying out the line, the rails were laid over it, if the hill 
proved too steep for this, the tracks went around it. As the 
road starts from sea level and reaches an elevation of 4,000 
feet in less than 100 miles, following a valley between moun- 
tains of 9,000 feet, there are many steep grades and sharp 
curves; in fact, for half the distance it becomes a cog road. 
But it is one of the most picturesque routes imaginable, 
running now through the green rice fields, now past fields 
ready for the sickle for there are no seasons, and sowing, 
reaping, and threshing may be under way in the same field 

now through a grove of cocoanut and banana palms, and 
again through a dense tropical jungle, where a dozen chatter- 
ing monkeys scared up by the train go swinging away from 
limb to limb. On our first trip over the road we had just 
come to a bridge and were going down a steep grade when 
the Malay brakeman in some way lost his headgear. While 
we sat there for half a second wondering what he would do 

for we had never seen a native without his topi he 
solved the difficulty by jumping off the train, running back 
after his cap, and after a short run of 50 yards, catching the 
train again, which all this time was moving at its usual 
speed. At one of the stations where we got off for a few 
minutes to look around, we were immediately the center of 
a large crowd of most curious natives. They offered us 
bananas pisang as they call them and oranges, mostly 
green in color, little pieces of sugar-cane on slivers of wood, 
and many strange and wonderful looking messes. They 
eyed us with much wonder, and attempted a few words in 
Malay, which we however understood about as well as they 


did our English. In a few days however, they became ac- 
quainted with our mission, and spoke of us by the Dutch word 
"Zoneclips," while we, on the other hand, learned to make 
ourselves understood in Malay. 

It was necessary to select the three Naval Observatory 
eclipse stations on the line of the railroad, on account of 
the difficulty of transporting heavy instruments and supplies 
into the interior. At Fort de Kock, near the edge of the 
moon's shadow path, three members under the direction of 
Professor W. S. Eichelberger, U. S. N., investigated the 
corona and the atmosphere of the sun with photographic 
telescope and spectroscope. At Solok, the main station of 
the party was located under the direction of Professor A. N. 
Skinner, U. S. N. Here was Professor Barnard with a pho- 
tographic telescope of 6i^-foot focus, which would give an 
image of the sun about seven inches in diameter. Barnard 
planned to give several short exposures, together with a single 
exposure of two and a half minutes, on a plate 40 inches 
square, hoping thereby to obtain some exquisite details of the 
corona. At Solok were also located several spectroscopes 
under the supervision of Jewell who was engaged in various 
interesting objects of research. 

Abbot of the Smithsonian Institution was investigating 
along two separate lines; first, with four telescopes mounted 
so as to photograph a large region in duplicate in the vicinity 
of the sun, he was attempting to discover new members of 
the solar system revolving inside of the orbit of the planet 
Mercury (if there be any such objects). His second task 
was a plan for measuring the heat of the corona and of the 
dark side of the moon turned towards us with the delicate 
bolometer, which is in reality nothing but a very sensitive 
thermometer capable of detecting the heat of an ordinary 
candle at the distance of five miles and of measuring dif- 
ferences of temperature to the one-millionth of a degree. 

The author was in charge of the U. S. Naval Observatory 
party at Sawah Loento at the terminus of the government 
railroad. The sawah, or irrigated rice-fields, is one of the 
prettiest sights imaginable. These fields are usually on the 
side of a hill, for the west coast of Sumatra, where the as- 


tronomers were located, is extremely rough and rugged. 
The sides of the hills have been terraced in order that the 
water may flow from one level to the one next lower, the 
enormous amount of manual labor required to make these 
terraces being for the most part performed by the women. 
The rice grows under water, but artificial irrigation is not 
needed, for it rains almost every day, not merely a shower, 
but with an average fall of half an inch for every day of the 
year. A view of these rice terraces with their vivid tropical 
greens and such brilliant color is never seen in the tem- 
perate zones is certainly a magnificent sight. All trace of 
the sawah, however, has departed from the village of Sawah 
Loento, and its place has been taken by the Ombilian Coal 
mines, belonging to and operated by the Dutch government, 
which are remarkable for having a vein of rich coal forty 
feet in thickness. 

The elevation of the railway station at Sawah Loento, 
according to the Dutch maps, is 262 meters above sea level, 
or 859 feet. The height of the eclipse station above this 
was measured by means of an aneroid barometer as 400 feet, 
which therefore made the elevation of the eclipse camp 
about 1,260 feet above sea level. A short distance away 
was the expedition from the Massachusetts Institute of 
Technology consisting of Burton, Hosmer, Harrison Smith 
and Matthes. 

The solar problems to be investigated by the Naval Ob- 
servatory party at Sawah Loento were along two separate 
lines: to photograph the corona with a camera 104 inches 
long, with lens six inches in diameter, and secondly, to 
photograph the spectrum of the chromosphere with a plane 
grating and quartz lens, used objectively without slit. To 
mount the instruments, piers of brick were constructed. 
Tents were used as coverings for the instruments, with an 
extra one for a storehouse, four tents altogether being set 
up. As the Boston party had reached Sawah Loento about 
ten days before us, they had learned to some extent the best 
way in which to proceed. The benefit of their experience 
was freely imparted, and much valuable information cheer- 
fully given. To Meinheer van Lessen, the chief engineer 


of the coal mines, many thanks were due for the very gen- 
erous way in which he looked after the supplying of all 
building material. The bricks, made near the coal mines, 
were transported to the eclipse location in three stages ; first 
by rail to the residence of the controller; second, by the 
slow-moving "kreta kerbau," drawn by the sturdy water- 
buffalo; and third, for the remainder of the distance, about 
a third of a mile, by coolies. The slowness of the last part 
of the operation can perhaps be imagined when it is stated 
that in a basket slung on a bamboo pole on the shoulders 
of two coolies, five ordinary sized bricks would be carried. 
This was indeed the minimum load, but the maximum was 
never more than ten. It was very interesting to see six coolies, 
using three bamboo poles, carrying a barrel of cement. The 
coolies provided were convicts. The way in which the Dutch 
East Indies treat their prisoners is one of the interesting fea- 
tures in the management of the Malay. If a native of Su- 
matra commits a crime, he is sent to one of the other East 
India islands ; and similarly natives of the other islands are 
sent to Sumatra to serve out the terms of their punishment. 
Consequently, in a penal settlement in Sumatra are natives of 
Java, Borneo, Celebes and of some of the smaller islands, but 
no Sumatrans. The reason for this strange separation is that 
there is great enmity between the different races, and so if a 
Javanese prisoner tries to escape he is immediately appre- 
hended by the first Sumatran who meets him and sent back. 
And thus it is that there are large penal settlements with no 
surrounding walls and very few guards. 

Such a settlement was at Sawah Loento. The convicts, to 
the number of 3 ,000, were employed in the coal mines. There 
was no guard over them except the " mandur " or policeman, 
one of themselves raised to a position of authority and 
responsibility, and answerable for the conduct of the 
coolies under him. Notwithstanding these conditions, one 
could go around with perfect safety and with less fear of 
molestation than in New York City. In fact, the pistols 
we had brought, with which to defend ourselves from canni- 
bals, were soon packed carefully away in the bottom of our 






In order to make the convicts more satisfied with their lot, 
the Dutch authorities paid them the amount of seven cents 
Dutch money, 2.8 cents American, per day. There seemed 
to be a great amount of sickness among the convicts, pos- 
sibly due to their confinement in the hot country, and several 
hundred of them were in hospitals or were put on sick duty 
without pay. These coolies were provided for us by the con- 
troller free of charge, and we were able to have about as 
many as we needed. They were very slow and not over fond 
of work, but nevertheless they could carry bricks about as fast 
as our Malay " tukang " could lay them. He used to squat 
at his work, and it took five days for him to lay 2,200 bricks. 
But the piers were finally built, the tents raised, and the in- 
struments gradually mounted. 

The day of the eclipse dawned clear, and our hopes were 
that these favorable conditions would remain until after 
totality, which occurred shortly after noon. First contact 
was observed in a perfectly clear sky, but soon after this, 
clouds began to gather, and a quarter of an hour before second 
contact the sky was completely overcast. 

The disappearing crescent of the sun was observed with 
a binocular before one barrel of which was arranged a small 
plane grating in such a way that with one eye the spectrum 
could be seen, and with the other eye the sun itself. With 
this, shortly before the time of second contact, bright lines 
were seen for a few seconds at F and H and in several places 
in the green and yellow, but these disappeared almost at 
the instant of being seen, the sun being completely hidden 
by clouds, and the flash passed without our being able to see it. 

Toward the middle of totality, conditions became a trifle 
better, so that it was possible to see, through clouds, the 
corona extending for about half a diameter from the sun, 
and with the small spectroscope to trace the form of the 
coronium line quite distinctly. During no time of the S m 41" 
of totality was an unclouded view of the corona obtained, 
but nevertheless, the clouds were so thin at the end of totality 
that the second flash was beautifully seen. 

One hour after the total phase the clouds cleared away 
and a perfect sky remained for the rest of the day. Alas! 


that the eclipse did not occur at one o'clock instead of 
twelve ! 

The pre-arranged program was carried out, however, in its 
entirety as if it had been clear. 

Powerful spectrographs were used also at the two other 
Naval Observatory stations. At the main station near Solok, 
Jewell had a concave grating, but as the result of the ex- 
perience in 1900, the instrument was used without a slit. 
Dinwiddie had a prismatic camera under his care, a single 
large prism of 60 angle, kindly loaned by the Smithsonian 
Institution. Near the edge of the moon's shadow-path at 
Fort de Kock, Humphreys had a huge concave grating of 
nine inches aperture, with lines three inches in length and a 
radius of curvature of thirty feet. This was the largest 
grating ever ruled on Rowland's dividing engine and it was 
constructed specially for the purpose of this eclipse. It was 
not a perfect specimen of a grating for the diamond point 
had broken down during the course of the ruling. This 
station was purposely located near the edge of the shadow 
in order to photograph the low-lying layers of the chromo- 
sphere. Weather statistics did not promise favorable skies 
at Fort de Kock but strange to relate, it was the only 
locality in Sumatra where astronomers were stationed that 
clouds were not experienced. The main station at Solok 
was far from being fortunate, and the astronomers there suc- 
ceeded in getting almost nothing, so dense were the clouds 
throughout the whole of totality. 

From Jewell's many years at Johns Hopkins University, 
working as assistant to Rowland, he had had vast experience 
in the development of spectroscopic plates, and hence the 
task of developing the precious negatives secured was handed 
over to him. While this was being done, and while the instru- 
ments were being dismounted and boxed and crated for their 
long journey home, we had some days to look around and 
observe the manners and customs of the people. The natives 
are all of them frightfully lazy or shall we say only tired? 
But it is the same disease that afflicts the natives all through 
the East. The Filipinos are troubled in the same manner, a 
similarity not to be wondered at, for Sumatrans and Filipinos 


both belong to the same great parent family. It was prob- 
ably for this reason that the natives of the Dutch East Indies 
were so interesting. 

The Malay is naturally a man of easy-going, indolent 
character, who never gives open expression to a sense of as- 
tonishment or fear, and is probably little affected by these 
feelings. When alone, he is gloomy and taciturn, never either 
singing or talking to himself. The upper classes are exceed- 
ingly courteous, yet this outward refinement, strange to say, 
co-exists with the most pitiless cruelty and contempt of hu- 
man life, traits which belong to the dark side of their char- 
acter ; and herein lies the explanation of the many diametri- 
cally opposed judgments which have been given us of the 
native of the East. There is another trait we must not for- 
get, and that is an insatiable love of gambling which no laws 
seem to be able to suppress. 

These are the characteristics of the Malay, but they de- 
scribe the Filipino just as accurately as they do the native 
of Sumatra. We were almost two months in Sumatra and 
in that time were thrown constantly with the natives who, 
as servants and coolies, were used continually. It became 
necessary to learn Malay and speak it ; this however proved 
not to be a very difficult task as the language contains neither 
declensions nor conjugations. 

The horned-roofed houses were very picturesque, and near 
Fort de Kock one of these was seen being modelled in gold 
and silver filigree work (at which the natives are very skill- 
ful), for a present from the colony to the Queen of Holland, 
to whom they are very loyal. The natives have a pretty 
legend, the Menangkarbau, about the origin of the horn- 
shaped roofs. This relates that there was once eternal 
enmity between the Javanese and Sumatrans, first one side 
conquering and then the other, and it seemed likely that the 
bloody conflict would continue without anything ever being 
decided. Finally the sages of the two peoples hit upon a 
plan to the effect that each nation should choose a champion 
from the animal world, that these should meet in mortal 
combat, and that whichever animal was victor, that nation 
was to be considered the conqueror for all time. The Java- 


nese chose a vigorous young tiger, under the impression that 
their champion surely could not be beaten, while the fore- 
fathers of the Sumatran natives picked a strong two-year- 
old karibou bull. Fancy, if we can, the two tribes in half 
circles meeting to see the fray, the shouts of rejoicing on one 
side and the woes of the other when the karibou succeeds in 
using its horns to such advantage that the tiger is killed. To 
perpetuate this victory, the Sumatrans forever after build 
their roofs to show the karibou horns. Usually the roof has 
one pair of horns, but it may have two or three pairs, the extra 
pairs, we are told, signifying that a daughter is married in the 
house. For the family in Sumatra is maternal, with the 
mother as head of the house, so when a daughter is married 
she brings her husband to live in her mother's home. Strange 
to relate, this form of family government coexists with Mo- 
hammedanism, where a man may have three or four wives. 
In such cases, however, we were never given to understand 
just how a man divided his time, and how he could live peace- 
ably with so many mothers-in-law. Women do most of the 
work, till the soil, gather the crops, thresh the rice and carry 
all the burdens, while the man evidently considers himself a 
superior being and succeeds in doing very little work. No 
wonder then he has three or four wives to work for him ! 

The Dutch have kept the original tribal relation and each 
tribe still has its chief, who is however, little more than a 
figurehead. Each man is required to perform one hundred 
days labor each year for the government, and as a great 
amount of this is put on the roads, excellent carriage roads 
appear even in almost inaccessible parts of the island. The 
principle under which the Dutch run their colonies has, by 
some writers, been described as merely for the purpose of 
making money out of them, and hence it has been the policy 
to keep the natives from becoming educated or Christianized, 
to keep out European immigrants and capitalists and to pre- 
serve the whole trade as a monopoly of the home government. 
And so we hear of the " culture system/' but the products 
under the " system " have gradually been reduced, until now 
in Sumatra the chief ones are coffee and tobacco and who 
has not heard of " Sumatra wrappers "? Natives are obliged 


to plant a certain number of coffee trees each year, and sell 
the product at a fixed rate to the government. The Dutch 
have, however, been exceedingly successful in keeping the 
Malays docile and contented, and it might be well for the 
Americans to study the results of these three hundred years 
experiment on Malays of exactly the same characteristics as 
the Malays in the Philippines, the Filipinos. 

After traveling half way round the world in search of 
knowledge, to have cloudy weather during the precious six 
minutes of the eclipse was indeed heart-rending, and the 
majority of the astronomers were pretty blue as a result. 
But when the plates came to be developed and what a 
boon the photographic plate has been to the astronomer 
it was found that the clouds had not interfered quite as 
much as expected. And so, down on the coast, Dyson and 
the Dutch parties, with the assistance of their warships, suc- 
ceeded in getting, with telescope and spectroscope, some 
really excellent photographs of the corona and " flash." The 
corona could not be traced a very great distance from the 
sun, but many exquisite details were seen in the inner corona. 
At Padang, Perrine of the Lick Observatory had the same 
kind of cloudy weather as experienced by nearly all the 
other scientists, yet the results of his work were very satis- 
factory in spite of the clouds. The photographs of the corona 
taken with the 4O-foot photoheliograph, like those of the 
English and Dutch parties, showed splendid detail in the inner 
corona, but they were particularly interesting from an ap- 
pearance in the northeasterly portion of the corona, as if an 
explosion had taken place. Perrine later found that this 
remarkable disturbance was immediately above the promi- 
nent and only sun-spot visible during eleven days, thus show- 
ing an intimate connection between sun-spots and disturb- 
ances in the sun. 

The three parties at Sawah Loento suffered likewise from 
the clouds. The main work of Burton and his party from 
the Massachusetts Institute of Technology was investigat- 
ing the magnetic disturbances while the eclipse was in prog- 
ress; this work, however, could be carried out as well in 
cloudy weather as in clear. The Boston party found the 


disturbance not so great as at the eclipse of 1900. Newall 
of Cambridge, England, had a very complete spectroscopic 
program to carry out, including many interesting researches, 
among which was an attempt to measure the velocity of rota- 
tion of the corona and to see whether or not this halo shines 
by its own inherent light. 

The few days previous to departure were spent at Padang 
where we had the pleasure of meeting many of the visiting 
astronomers whose accommodation taxed to the utmost the 
capacity of the leading hotels, the Oranje and the Atjeh. On 
the evening of May 27, the United States consular agent gave 
a dinner and farewell ball in honor of the American astrono- 
mers and naval officers. Many prominent people were pres- 
ent, including Governor Joekes, also the officers of H. M. S. 
Pygmy. After the ball was over, the author slept in one of the 
beds of the Oranje Hotel with Professor Barnard, Dr. Abbot 
and one other. Our ship, the U. S. S. General Alava, had 
waited in Emma Haven to carry us back to Manila and we 
sailed at 10 A.M. May 28 on our homeward voyage. The 
eclipse of 1901 was a thing of the past! On June 8 we en- 
tered Manila Bay through the north channel, the Boca Chica, 
and anchored off Cavite. After a fortnight's wait in Manila, 
we sailed aboard the U. S. Army Transport Indiana, and 
reached San Francisco on July 16, five months from the day 
we had departed. 

What scientific results had been obtained from the long 
journey and the many months spent away from regular 
duties? When the photographic plates were studied in de- 
tail it was found that much of great value had been secured 
in spite of the clouds, and as a result the Sumatra eclipse 
had contributed some very important information regarding 
the problems of solar physics. 



next eclipse to be widely observed was that of 
August 30, 1905. On this day the eclipse began 
in Manitoba and, after crossing through Northern 
Canada, it left Labrador about 8 A.M. on its trip across the 
Atlantic. Shortly after noon the shadow cut into Spain, 
then on through the Mediterranean, Northern Africa and 
Egypt, leaving the earth's surface at sunset on the coast 
of the Indian Ocean. 

Spain was chosen by the majority of astronomers, both 
because the duration of totality was longer, and because the 
promise of good weather conditions was better; and here in 
a path one hundred and twenty miles in width running diago- 
nally across the peninsula, hundreds of astronomers, Ameri- 
can and European, were gathered. 

The party sent out by the United States government was 
under the general direction of Rear-Admiral Colby M. 
Chester, U. S. N., superintendent of the Naval Observatory. 
Three men-of-war were furnished by the Navy Department 
for the purposes of the expedition, the U. S. S. Minneapolis, 
U. S. S. Dixie and U. S. S. Caesar, the first named being the 
flagship of the squadron. 

The three vessels left separately from the United States 
about the end of June and met in Gibraltar about the middle 
of July. " Gib " is one of the most interesting places in the 
world, especially when entering on a naval vessel. It was 
a glorious sight, as we steamed in at dawn on board the 
Minneapolis, to behold the wonderful rock and, sheltered at 
its base, the Mediterranean squadron of the British navy, 
consisting of eight battleships and eight first-class cruisers, 
under the greatest of English admirals of the time, Lord 



Charles Beresford. The morning of our arrival was spent 
in firing and acknowledging thunderous salutes, and in mak- 
ing official calls. To carry out properly these acts of courtesy 
between the American and British nations, it was necessary 
to fire no less than one hundred and fifty-two rounds of am- 
munition. On the morning of our second day in Gibraltar, 
the British squadron sailed, and it gave us an idea of the 
quality of the greatest navy in the world to see the splendid, 
seamanlike manner in which the big ships got under way 
without confusion, and one by one in perfect order departed 
from the crowded harbor. 

After leaving Gibraltar and entering the blue waters of 
the Mediterranean, the Minneapolis steamed along the coast 
of Spain for about four hundred miles and anchored in the 
harbor of Valencia, the first American man-of-war to visit a 
Spanish port since the Spanish-American War. 

It had been decided to divide the Naval Observatory ex- 
pedition into three, sending two parties to Spain and one to 
Africa. In Spain the parties were located, one at the edge of 
the path of totality at Puerto Coeli, and the other near the 
central line at Daroca. 

Daroca is in the heart of old Spain, about forty miles 
from Saragossa, and as a railroad had been there only four 
years it was a terra incognita for modern tourists for 
which we were duly thankful. Our six weeks' stay there 
was a happy commingling of hard work and there was 
plenty of work to do with pleasant experiences in getting 
acquainted with Spanish life and people. The site for the 
town is indeed a peculiar one, in a valley so surrounded by 
hills that each heavy rain storm used to flood the city, until 
about 1600, a tunnel was constructed through one of the 
hills to carry away the waters. The tops of these hills are 
crowned with walls and forts, most of them constructed by 
the Moors a thousand years ago, some of them by the Catho- 
lic Spanish since that time. There is one tower of special 
interest, and still in good state of preservation, which is said 
to have been built by the Romans before Saguntum was 
founded, and it is, therefore, more than two thousand years 
old. (The railroad from Valencia passes through Saguntum 







where Hannibal and the Romans had their memorable fight 
in B.C. 238.) 

The Spaniards received us with open arms and did every- 
thing in their power to assist in our work and to make our 
stay in their midst as pleasant as possible. As no one in the 
place could speak English, it was necessary to make ourselves 
understood in their language. 

To help in the erection of the temporary observatory, six 
sailors were sent in from the Minneapolis, and all hands, 
astronomers and sailors, worked each day from early morn- 
ing till late at night, building piers, erecting telescopes with 
houses to shelter them, mounting spectroscopes, and fixing 
up a meteorological observatory. After the carpenters and 
machinists had finished their work of construction, it was 
necessary for the scientists to focus and adjust, to see that 
everything was in good working order, and to make trial 
photographs. A few days before the eclipse, the party was 
increased in size to thirty-five, officers and sailors having 
come up from the ship for the purpose of assisting in the 
observations. Frequent drills were held in order to fa- 
miliarize each one with his part and thus to be sure that 
everything would go right and that no precious seconds would 
be wasted at the time of eclipse. 

The location of the eclipse camp was half a mile south of 
the town, in the midst of a beautiful, fertile valley. From 
there, while we worked, we could catch glimpses of scenery 
typical of Spain. The first feature to attract one's atten- 
tion is the extremely barren aspect of the country, which 
is in sharp contrast with the garden-like appearance of Eng- 
land. The hills of Spain were in early times densely wooded, 
but now are almost entirely devoid of trees and look from 
a distance as if there were not a particle of vegetation on 
them. Moreover, the rainfall is so slight that agricultural 
pursuits must rely upon irrigation, and thus it is only the val- 
leys that are green and cultivated. In such a valley along 
the shores of the little river was our eclipse camp located. 
The greenest field was decided upon as the site of the observa- 
tory, and upon application to its owner for permission we 
found that he was quite satisfied to allow his plot of ground 


to be used, but thought some compensation should be made 
for the valuable crop of grass that might possibly be raised 
during the summer. On receipt of one hundred pesetas, he 
forthwith proceeded to take a fatherly interest in all of our 
doing,s, and explained scientific matters to every one as if he 
had been chief of the expedition. His field became the cen- 
ter of interest in the community, and people came from all 
sides to look upon the strange doings. As a prominent trait 
of the Spanish peasant seems to be a great and overpowering 
curiosity, we had pleny of onlookers; and when the mayor 
and a few of the most prominent citizens were invited to look 
at the moon through our five-inch telescope, we were rather 
surprised to put it mildly to find over one hundred peo- 
ple turn up, when only a half score had been invited. Their 
curiosity took the form only of making each and every one in 
the town intensely interested in what was going on, and to 
show that interest they turned out in force each afternoon to 
see how matters were progressing. It might be asked, what 
their attitude was towards these Americans who had so 
lately beaten them in the small war. Before the expeditions 
reached Spain, it was feared that perhaps there might be 
some friction on that account, but these fears were not real- 
ized. As a matter of fact, the only person we met who 
seemed to have any feeling at all in the matter was a former 
soldier in the Spanish army. He had seen service in the 
Philippines, had been captured and thrust into prison by the 
Filipinos, had been rescued by the Americans, and as a re- 
sult he had only the kindest feelings towards everything 
belonging to the United States. As for the rest of the peo- 
ple, they seemed to have forgotten all about it, or else they 
did not know there had been a war, for it must not be for- 
gotten that only about one quarter of the people in Spain can 
read and write. 

After this my third total eclipse, I can confidently say that 
observations at such a time consist of much hard work and 
many nerve-racking experiences. The astronomer is never 
on hand sufficiently long beforehand to take things quietly 
and easily, he must work under conditions to which he is 
totally unused, and over his head hangs the knowledge that 


everything must be completed by a certain day and a certain 
hour, for the eclipse cannot be postponed, and there is no sec- 
ond trial in case of failure. In addition to working hard all 
day as carpenter and instrument maker, the astronomer must 
stay up half the night adjusting his instruments on stars, so 
that during the last few days before the eclipse very few 
hours of sleep each night are obtained. However, in spite of 
the many difficulties that were continually cropping up, the 
mounting and adjusting of the instruments was practically 
completed by August 25, when our observing party was 
swelled in numbers by the officers and men from the Min- 
neapolis. From then until eclipse day the time was spent in 
putting the finishing touches on the work of adjustment, and 
in having frequent drills in order to insure that everything 
would go without a hitch. 

What was the promise of weather for the important day? 
We had been closely scrutinizing the weather each day to 
see what conditions we were to expect, and were much 
pleased to find that the sky was usually clear just after noon, 
the hour when the eclipse was to occur. August 29, how- 
ever, had been cloudy all day so that on eclipse day we had 
to go to camp early to test our final adjustments, go through 
drills once more and to be sure that all the apparatus worked 
smoothly. The skies were clear and our hopes for success 
were high. Outside the roped-off enclosure, the whole town 
of Daroca was assembled, for it was naturally thought by 
the people that nowhere could the eclipse be seen so well 
as where the astronomers were located. 

At 11:52 A.M. a little indentation was seen on the western 
limb of the sun, and the eclipse had begun. The skies were 
clear with the exception of a cloud here and there, and our 
most ardent wish was that the clouds would leave the sun 
clear for the next couple of hours. For the first hour that 
the moon was creeping over the sun there was nothing of 
very great moment to notice, but for the next twenty min- 
utes till 1:12, when the sun was blotted out, we were each 
of us filled with expectancy, for matters began to take on 
a weird and unnatural appearance. The little blotches of 
light under the trees, instead of being familiar circles, were 


little crescents, exact counterparts of the sun itself. The 
darkness began to make itself really felt, and without look- 
ing at the sun one would know that something out of the 
ordinary was happening, for the gloom did not in the least 
resemble that of sunset. A hush fell upon the crowd of as- 
sembled and talkative Spaniards when, ten minutes before 
totality, a big cloud drifted over the sun. Would this cloud 
move away? Or were we going to be disappointed? It 
hung there for a space of time that seemed to be an age, 
while in reality it was only five minutes. It was a big scare, 
but when that passed, with a shout from us all, there wasn't 
another cloud anywhere to bother us. Seventeen seconds 
before the calculated time, with the last disappearing ray of 
sunlight, the corona broke forth into view. What a mag- 
nificent sight it was shining out with its pale, pearly light 
for a couple of diameters round the edge of the sun, with 
its streamers and brushes of delicate light! True to pre- 
diction, the corona was almost square in shape, and was not 
at all alike in appearance to the other coronas the writer had 
seen in 1900 and 1901, with their long fish-tail extensions 
along the sun's equator and short-curved streamers near the 
sun's poles. In the upper left hand quadrant, huge red flames 
sixty thousand miles high could be seen with the naked eye, 
and these with a closer view with the telescope resolved them- 
selves into a forest-like structure. Close to the sun the corona 
was very bright, in fact so brilliant that the eye was not read- 
ily able to take in all the details of the faint streamers. As a 
pictorial effect, without the long equatorial extensions, this 
corona was much inferior to the two last ones seen. Still it 
was a magnificent sight, and we were more than thankful for 
having clear skies for making our observations. 

When totality first started we were each and all of us much 
too busy to take very much notice of our immediate sur- 
roundings or even of the corona itself. We could not help 
becoming aware that our Spanish onlookers outside the ropes 
were appreciating the show in the skies provided for them 
without expense. From the noise made each one seemed to 
be telling his neighbor at the top of his voice just how it hap- 
pened and what there was worth seeing, and this in spite of 


the fact that the mayor of Daroca had generously provided 
half a dozen members of the civil guard to preserve order and 
keep quiet. For the first half minute the din was so great 
that it was impossible to hear the seconds counted, or to know 
exactly when to begin and end the exposures of the photo- 
graphs. When the Spaniards had quieted down, after their 
first outburst, all that was heard in the eclipse camp was the 
steady count of the observer calling out the seconds as they 
passed, the quiet words of the observers giving commands to 
their assistants and the click, click of the various pieces of 
apparatus as exposures were made and plate holders removed. 
Everything passed off without a hitch, and with the first re- 
appearance of the sun our work was over and we could take a 
long breath. 

We had been favored with clear skies. How many others 
were equally fortunate? It did not take us long to find out, 
for the Spanish government had installed right in our camp 
a telegraph office, and for fifteen days no less than three 
operators were at our service to send and receive our mes- 
sages; and for this not a single cent of money was asked 
or expected. It was found that fifty miles to the west of 
us, at Alhama, where were the observers from the Lick Ob- 
servatory under Professor Campbell, there were thin clouds, 
while one hundred miles to the east along the Mediterranean 
coast, the Englishmen were even more unfortunate in hav- 
ing the clouds denser. In the northeastern part of Spain at 
Burgos, more astronomers were located than at any one place, 
and here too was King Alfonso of Spain. Five minutes before 
totality it was pouring rain, but as if by a miracle a little blue 
patch of sky appeared, and the eclipse was seen under per- 
fect conditions. The weather along the eclipse track was: 
in Labrador, cloudy (no observations made) ; in Spain, cloudy 
and clear; in the islands of the Mediterranean, cloudy; on 
the coast of Africa, slightly cloudy; but farther inland and 
along the rest of the eclipse track the skies were perfect. All 
three parties of the Naval Observatory were fortunate in hav- 
ing their work unhindered by a single cloud. 

My own work was entirely spectroscopic. The photo- 
graphic plates were developed within the walls of the col- 


lege of Daroca, and in the long hours necessary for this work 
I was greatly encouraged and assisted by my good friend 
the rector of the college, Padre Felix Alvirez. Daily inter- 
course with this reverend father endeared him to me very 
much; and Senors Lorente, Soria and Padre Felix made my 
stay in Daroca one of the most interesting spots of my whole 
life by the kindness with which they bore my imperfect Span- 
ish, by the interesting bits of history they told of Daroca, and 
by the deep insight each gave into the courtesy of a Spanish 
gentleman's heart. The developed photographs of the flash 
spectrum showed exquisite definition over a wide range in 
wave-lengths. Evidently I had stored up for myself the ma- 
terial for months and years of intensive study, the results of 
which will be told in Chapter XVI. 




HIRTEEN years after the splendid observing con- 
ditions of 1905 in Spain I witnessed my fourth total 
eclipse of the sun, again as a member of the ex- 
pedition representing the United States Naval Observatory. 
On June 8, 1918, the shadow of the moon touched the earth's 
surface on the Pacific Ocean, far south of Japan. Owing to 
the revolution of the moon about the earth and to the rota- 
tion of the earth on its axis, the shadow crossed the Pacific 
Ocean at a speed of over a thousand miles per hour. It was 
well after noon before the shadow reached the American con- 
tinent, and the eclipse began in the state of Washington. 
Here the width of the shadow was only sixty miles, so that 
only those fortunate enough to be within this narrow track 
were able to see the eclipse in its totality. The eclipse passed 
southeasterly through Washington, Oregon, Idaho, Wyo- 
ming and Colorado in succession. In Colorado, the shadow 
had dwindled to forty miles in width. After passing through 
some of the central states, the shadow left the United States 
at Florida and left the earth's surface in the Atlantic, off the 
coast of the Bahama Islands. 

The eclipse was seen almost exclusively from the United 
States, and so it will be known as the American Eclipse of 
1918. Since more than half the civilized world was in the 
grip of the tremendous war, it was necessary for American 
astronomers in the year 1916 and early in 1917 to make 
their plans to insure that this eclipse should be well ob- 
served. Before our own country had become involved in 
the war, Congress had been asked for, and had made, a 
special appropriation to defray the expense of equipment 
and travel for the party from the U. S. Naval Observatory. 

In order to help the astronomers of the country to make 



as intelligent a choice of an eclipse site as possible, the Naval 
Observatory, in 1917, had prepared a large scale map of the 
United States showing, among other things, railroad lines, 
contour lines and the location of towns within the eclipse 
track. The city of Baker in eastern Oregon seemed to be the 
ideal spot for the government party. The question of clear 
skies was the all-important one for the proper location of an 
eclipse party, but, fortunately, the U. S. Weather Bureau 
had a regularly equipped station at Baker, and a record of 
many years 7 continuous observations seemed to be the ideal 
method of securing the desired information regarding the 
probabilities of good conditions on the day of June 8. As 
the Weather Bureau promised an absence of clouds and rain 
with an abundance of clear skies, Baker was chosen with the 
great hope that it would live up to its good reputation in the 
matter of weather. 

We were in Baker exactly six weeks before eclipse day, 
and the time was none too long. The apparatus was sent for- 
ward by through freight, and though we greatly feared de- 
lays, it arrived safely the second day after our own arrival. 
To assist in the work of erecting the apparatus, the superin- 
tendent of the Naval Observatory had requested the services 
of five sailors from the U. S. Naval Station at Bremerton, 
Washington, who were in charge of a chief petty officer. The 
sailors were carpenters and machinists who assisted the as- 
tronomers in splendid style so that ten days before the eclipse, 
when the balance of the party began to arrive, the apparatus 
was all erected and partially adjusted, and there remained 
only the perfecting of the adjustments in order to be ready 
for the all-important day of the eclipse. 

For direct photography of the corona, the largest camera 
was one of sixty-five feet focal length arranged horizontally, 
the light from the eclipsed sun being reflected by a coelostat 
mirror. The spectroscopic work of the Naval Observatory 
party called for the use of three concave gratings, each used 
objectively without slit. The largest instrument was a 
twenty-one-foot Rowland grating of six inches aperture and 
15,000 lines per inch. This had a spectrum specially bright 
in the first order on one side, the grating being kindly loaned 


by Professor J. S. Ames. Photographic films two by twenty- 
four inches were used, and it was planned to work from the 
extreme ultra-violet as far to the red as the length of the 
films would permit. The second concave grating was of ten 
feet radius and 15,000 lines per inch, the grating belonging 
to the Naval Observatory. This was used in the first order 
and gave the same dispersion as the instrument employed in 
Daroca, Spain, in 1905. By the use of special emulsions 
kindly prepared by Dr. C. E. K. Mees of the Eastman Kodak 
Company, an attempt was made to photograph farther to the 
red end than in 1905. The third one was a very short focus 
grating of two meters radius and of six inches aperture be- 
longing to the Astrophysical Observatory of the Smithsonian 
Institution. This grating gave very brilliant spectra but 
with little dispersion. The films for this spectrograph were 
stained with dicyanine. 

Fortunately for the work of preparation, and true to the 
prediction of the U. S. Weather Bureau, no rain fell during 
the entire stay of the astronomical party in Baker. Accord- 
ing to the " oldest inhabitant/' the season was unusually dry 
even for eastern Oregon. By some mysterious force un- 
known to the astronomers, the eclipse seemed to exert some 
potent influence over the weather. At any rate, it was as- 
serted by many of the rural papers that no rain could be 
expected until the eclipse was over. But if an absence of 
rain was experienced there was no lack of clouds, nor were 
the clear skies we had been led to expect afforded us. As 
the time for the eclipse drew nearer, the continued appear- 
ance of clouds began to cause anxiety among us. Would 
they interfere with the eclipse, and at the last moment make 
all the weeks of careful preparation of no account? If this 
had indeed happened, it would not have been the first event 
of the kind. Unfortunately for the astronomer, his work 
is always at the mercy of the clouds and the weather. But 
to have the whole work fail through clouds at the time of 
the few precious minutes of the total eclipse that is in- 
deed, the keenest sort of disappointment! Some astronomers 
seem to be always unlucky and always experience cloudy 
weather on their eclipse expeditions, while on the other hand 


others are always lucky, and sometimes after all hope is 
abandoned, a rift will appear in the clouds and the eclipse 
at totality be seen in all its glory. Would we at Baker be 
lucky or unlucky, would the clouds interfere or not? Nearly 
all the days spent in Baker, according to the classification 
of the U. S. Weather Bureau, were actually clear. A " clear " 
day is not necessarily cloudless from morning till night, but 
rather one when the " sky averages three-tenths or less ob- 
scured, from sunrise to sunset." Clouds, however, gathered 
almost every day shortly after noon, and this condition was 
usually accompanied by very high winds that at times rose 
to the strength of a mild gale. The eclipse was to occur dur- 
ing the middle of the afternoon, and at this time of day the 
skies were generally overcast. These same conditions pre- 
vailed over the whole of the western United States along the 
path where the astronomers were located. It was well to be 
an optimist under such conditions of sky, for the pessimist 
became more and more wretched as the day of the eclipse 
drew near and his law of averages showed him the almost 
certain chance of thinly clouded sky during the total 

Fortunately, so far in my eclipse experiences I had been 
among the lucky astronomers. In 1900, at my first eclipse, 
the weather was ideal not a single cloud in the whole 
sky. In 1901, 1 was a member of a rather large party which 
traveled half way round the world, of which only four of a 
total of thirteen saw the eclipse, the other nine witnessing the 
eclipse eclipsed by clouds. I was one of the fortunate four. 
Again, in 1905, there were many clouds which spoiled the 
researches of many parties. At Daroca, in Spain, a few min- 
utes before totality a dense cloud covered the sun, but it 
cleared away before the all-important time and the total 
phase was seen through a brilliantly clear sky. Three lucky 
chances out of three made a fine average. The hope was 
that June 8 would make it four out of four! 

By May 30, the whole party had assembled in Baker. A 
full week was given up to the final adjustments, and to the 
drills that were to play such an essential part in the work 
on eclipse day. During the partial phases of the eclipse, 


very few observations of importance were to be made; all 
observations of value came during the period of totality 
which lasted for one hundred and twelve brief seconds. If 
a slide of a plate holder should stick in place so that it could 
not be removed, or a lens were not uncapped at the proper 
time so as to let in the light, the whole work of an instru- 
ment might come to naught. On each day of the week pre- 
ceding June 8, drills were gone through several times, in 
the morning and again in the afternoon. These drills were 
so well carried out that on eclipse day each and every one 
performed excellently the task allotted to him with the re- 
sult that everything passed off without a single hitch. 

As the days in June progressed towards the eighth, there 
was an air of excitement as each astronomer grew more 
keyed up to the task before him. Would Saturday be clear? 
But more especially, would the two minutes from 4:04 to 
4:06 P.M. be clear? The skies were anxiously watched dur- 
ing the last days, but alas ! almost every day at eclipse hour 
they were overcast. The optimist reasoned that if it were 
cloudy all the days before June 8, then on eclipse day per- 
fect weather would surely be forthcoming; while the pessi- 
mist on the other hand argued that so many cloudy days 
meant still one more of the same character, so there would 
be no use trying to do anything. 

Saturday, June 8, dawned with the sky overcast with thin, 
filmy clouds. The sun was well visible through these clouds, 
however, and it was possible to examine again the focus that 
had been obtained with the spectroscopes and with a touch 
here and a touch there to decide that everything was in 
perfect condition. During the morning the drills were again 
practised, and these seemed to promise success. The weather 
during the six weeks had not held up the work, and every- 
thing that thought and work could do seemed now to have 
been accomplished. The astronomers who had been on the 
ground for the whole six weeks of preparation had the pleas- 
ant consciousness that all of their allotted tasks had been 
completed, that every little detail had been thought of and 
that perfect success would certainly crown their efforts if only 
the clouds would clear away. But during the course of the 


morning/ the clouds grew thicker instead of thinner, and it 
did indeed seem as if there was little chance of a clear sky. 

The first contact was to take place at 2:30 P.M. Shortly 
after noon, the city of Baker took upon itself the aspect of 
a holiday. Though the day was Saturday, all stores were 
closed from three until five in the afternoon so that every- 
one should have a chance to see the phenomenon. Naturally 
everyone in Baker wished to go to the eclipse site at the Fair 
Grounds, to watch the astronomers at work. At the eclipse 
in Spain, this had been permitted with the result that the 
whole town had assembled, each inhabitant jostling his 
neighbor to get as close as possible, and each apparently 
talking at the top of his lungs, with the result that such a 
din arose when the eclipse became total that it was impos- 
sible to hear the seconds counted off to give warning to the 
astronomers when to change their plate holders. 

In order that this might not happen again, the residents 
of Baker were told that the gates of the Fair Grounds would 
be closed, and that absolutely no one would be admitted 
within the enclosure, and the mayor of the city sent a guard 
of Boy Scouts to see that these orders were obeyed. Most 
of the townspeople repaired to the hills to the southeast of 
the city from which there could be obtained a fine view of 
the valley and the Elkhorn range, and they were directed 
to look especially for the shadow of the moon which would 
come across the landscape at the speed of thirty miles a 
minute or 1800 miles per hour. 

No appreciable improvement in the skies was observed 
from noon to the time of first contact. Through a thin patch 
in the clouds, Mr. Hammond, using the five-inch visual tele- 
scope, observed the beginning of the eclipse and made a 
record of it. The clouds if anything became thicker after 
this so that at three o'clock it was impossible to see even 
where the sun was. Little thin rifts appeared at times, so 
that it was possible to see the moon encroaching on the face 
of the sun. At three thirty, a patch of brilliantly blue sky 
was seen off to the northwest and as the precious minutes 
dragged along it became evident that the clouds were moving 
in such a way that it was quite possible that the blue patch 


would reach the sun in time for totality. Fifteen minutes 
before the total phase the clouds were so dense that had 
totality occurred then, the scientific results would have been 
nothing; but the blue sky was coming nearer and it might 
arrive in time. 

Without looking at the sky, one realized that something 
unusual was happening. The light of the sun became so fee- 
ble that even the birds felt the unnatural aspect of things 
and sang their songs as if they were going to rest. The cocks 
in the nearby farm crowed. The wind, which was ordinarily 
blowing at this hour, was quiet. All nature was hushed. 
Even the seasoned astronomers who had seen two or three 
eclipses before felt the thrill of the unusual spectacle. And 
still the question was, would the clouds clear away in time? 

At five minutes before totality the warning signal was 
given by Chief Petty Officer Welsch of the U. S. Navy who 
was to watch the chronometer and count the seconds. This 
signal summoned each man to his post. One last look was 
given to the apparatus to see that everything was in place, 
the plate holders were adjusted, and then we waited. 
" Two minutes " before was called out, and then " one min- 
ute/' still again " thirty seconds " before the expected time of 
totality. The clouds by this time had thinned considerably, 
the patch of blue sky was only a short distance away. The 
plan had been that after the signal of " thirty seconds," there 
should be nothing said until the word " Go " told that the 
total eclipse had begun. I was to watch for this with a pair 
of binoculars, before one glass of which a direct vision spec- 
troscope had been arranged. This was the plan followed in 
Spain with complete success. But due to the thin clouds at 
the beginning, it was impossible to see the spectrum lines 
with the spectroscope, and the signal " Go " was actually 
given by Mr. Hammond who was using the five-inch tele- 
scope. No sounds disturbed the work of the party except 
the call of the seconds as the time passed and the brief words 
of command and shift of plate holders as each member of the 
party did his allotted task. Ten seconds after totality com- 
menced, the clouds, thin at the beginning, had still further 
thinned, and at mid- totality the conditions were even further 


improved. What a gorgeous spectacle then met the eye! 
The sun was now in a very thin wisp of cloud with blue sky 
on either side. Although the cloud would undoubtedly de- 
tract from the scientific results, still it greatly enhanced the 
pictorial effect. The corona could be seen stretching for a 
short distance from the sun's edge, but most remarkable of 
all were three great tongues of flame, one immediately at the 
top of the sun, one on the left hand edge, and still a larger 
one on the right edge of the sun. These shone with a brilliant 
scarlet light, and made the eclipse of 1918 memorable as the 
eclipse of color. As the end of totality approached the thin 
clouds became still thinner, and two minutes after the 
eclipse was over the sun had reached the blue patch of sky. 
If the eclipse had occurred only two minutes later, or if the 
party had been only half a mile to the northwest, the sky 
conditions would have been perfect! If, as I have already 
said, the eclipse had taken place fifteen minutes earlier, the 
scientific results would have been nothing at all. The op- 
timists had won out. 

We had indeed been fortunate. But farther west at 
Goldendale, Washington, where the Lick-Crocker party was 
located, a change of weather had happened which amounted 
almost to a miracle. The account by Professor Campbell 
runs as follows: " The total phase of the eclipse occurred 
at 2:57, local mean time. By great good luck a small rift 
in the clouds formed mostly at the right place and right 
time. The clouds uncovered the sun and its immediate sur- 
roundings less than a minute before totality became com- 
plete, and the clouds again covered the sun less than one 
minute after the total phase had passed. The small clear 
area was very blue and the atmosphere was tranquil." * 

The developed photographs exhibit the painstaking care 
of the astronomers in procuring the precise focus, with the 
result that all of the photographs show exquisite definition. 
The thin clouds did not interfere at all with the details of 
the prominences or flames surrounding the sun. Those 
taken with the sixty-five-foot camera exhibit the promi- 
nences in splendid detail on a scale where the sun is more 

1 Lick Observatory Bulletin, 10, 2, 1918. 


than seven inches in diameter. The longer exposures for 
procuring the extensions of the corona were not quite so suc- 
cessful, since the thin, fleecy clouds cut down the fainter 
streams of coronal light. The smaller cameras showed the 
same results as the larger ones splendid detail in the in- 
ner corona, but the corona not of very great extent. All 
the photographs unite in showing many polar rays, and they 
also exhibit some plumed arches of great beauty. The corona 
appeared to be of the sun-spot maximum type, but with more 
polar streamers than were expected. 

The spectroscopes procured photographs of exquisite defi- 
nition, but these photographs suffered greatly owing to clouds 
which cut down the amount of exposure that at best is none 
too great. 

What was perhaps the most interesting piece of scientific 
work accomplished at the 1918 eclipse owes its conception 
to Mr. Edward D. Adams, of New York, who has shown his 
great interest in science by the founding of the Ernest Kemp- 
ton Adams fellowship which is awarded each year by Co- 
lumbia University for researches in the domain of pure sci- 
ence. Upon becoming a member of the United States Naval 
Observatory party, Mr. Adams took upon himself the re- 
sponsibility of trying, by some method, by photography, by 
a drawing, or by a painting, to procure a reproduction which 
would show the beauties of the corona, and which should be 
true not only as to form but more especially as to color. Un- 
fortunately for science, it is impossible to obtain a satisfac- 
tory representation of the corona and the sun's surround- 
ings by photography. The corona is very brilliant near the 
edge of the sun, but the intensity fades very rapidly. The 
eye can take cognizance of the details in spite of the great 
changes in brilliance, but not so the photographic plate. To 
obtain the faint extensions of the corona which are readily 
visible to the naked eye, a comparatively long exposure is 
necessary. This long exposure causes so much overexposure 
in the brighter inner regions of the corona that all detail 
there is lost by being burnt out. Short exposures give us the 
inner corona in exquisite detail, but the outer corona is then 
lost through shortness of exposure. Many attempts have 


been made to cut down the relative exposure by means of 
mechanical devices but none of these have been entirely 
successful. Heretofore, the only success in representing the 
corona has been obtained by taking photographs with differ- 
ent times of exposure and with different cameras in order to 
procure photographs with detail both in the inner and brighter 
parts of the corona, and in the fainter outlying portions. 
After the eclipse is over, a composite drawing is usually made 
from the examination of different photographs. This method 
has given several satisfactory drawings, but they still have 
left much to be desired. However perfect they may have 
been as drawings, they took no note of color. Mr. Adams 
took upon himself the task of finding the right man to draw 
and paint the corona. Color photography could not help out 
in procuring the right color, and there was left only the pos- 
sibility of finding an artist who would have the true scientific 
spirit, and who could combine an accurate sense of form with 
a refined perception of color. Mr. Adams was successful in 
finding Mr. Howard Russell Butler, a portrait painter of 
note, who has developed a shorthand method of noting both 
form and color. 

During the eclipse, Mr. Butler sat on a lofty perch over- 
looking the eclipse instruments, and from which he could 
obtain a fine view of the sun. The task he had taken to 
himself was no small one. As a portrait painter he usually 
asked for ten or twelve sittings of two hours each: now he was 
asked to render his subject in 112 seconds. And moreover 
this was the first corona he had ever seen ! 

The methods followed by Mr. Butler in painting the co- 
rona have been described by him in Natural History, IQ, 
244-271, 1919, and reprinted as Vol. II, part 6, of the Pub- 
lications of the Leander McCormick Observatory. An ab- 
breviated summary of his description is herewith given : 

" The method of working finally adopted may be called a 
shorthand method. It was to have a sheet of white card- 
board on the easel with a series of concentric circles and radii 
drawn upon it in advance. One of these circles was to have 
the same diameter as the photographs of the moon to be 
taken in the sixty-five-foot camera, namely, seven and three- 


eighths inches. There was to be an inner circle of half this 
diameter and outer circles whose diameters were respectively 
one and one half, two, and two and one half times that of the 
inner circle. I expected to use the seven and three-eighths 
inch diameter, and did actually use it, but I was thus pre- 
pared, in case of an unexpectedly extended corona, to reduce 
the scale to one half and get everything on the cardboard. In 
front and beneath my cardboard was a finished sample pic- 
ture of a corona, painted in advance as I expected it would 
appear, and my plan was to indicate by initials at points on 
my cardboard the variations of color from this picture ; thus 
b was to mean a variation toward blue from the sample pic- 
ture, and y more toward yellow. I wrote out the procedure 
as follows and tacked it alongside the easel. Practice enabled 
me to allot a certain number of seconds to each item. 

Procedure Seconds 

Note value and color of sky 10 

Draw value line on moon i o 

Note colors of moon 10 

Draw outline of corona 20 

Use Zeiss binoculars 20 

Record positions of prominences 10 

Note color and value of prominences 10 

Note colors and values of corona, etc 20 


" Then my plan was to paint a first picture from this result- 
ing memorandum, while the impression was vivid, and as soon 
as there was sufficient light to proceed by. 

" While disappointed in not seeing the corona in a cloud- 
less sky, the thin veil had its advantage from the artist's 
standpoint. It added mystery and the effect was picturesque. 
The brilliant corona burned through the thin veil as if it were 
not there. Probably only the outside edges of the corona 
were affected. 

" On the tenth, the photograph negatives were shown to 
me. Those of the sixty-five-foot camera were seven and 
three-eighths inches in diameter, the others considerably 
smaller. I now saw, in minute detail, the two prominences 
which I had recorded and the mighty cyclone which had 


been increasingly revealed as the eclipse neared its end, be- 
cause of the direction of the moon's motion. There were 
many other minor prominences. 

" I now made careful drawings of these prominences from 
the negatives and of the variations in shading of the sur- 
rounding corona. Many arches were found springing over 
the prominences, and a few rifts of dark channels radiating 
from the limb but never coming very close to it. The nega- 
tives showed very clearly the hairy polar rays, not always 
radial in direction, and the beginning of a wing springing 
from the upper right-hand limb of the sun. 

" Three paintings were made, the first immediately after 
the eclipse, the second on the succeeding day and the third 
after all data had been secured. This final painting is the 
one reproduced on page 60." The original size of the paint- 
ing is 49 x 33^- inches. 

One of Mr. Butler's paintings has been presented to the 
American Museum of Natural History in New York by Mr. 
Edward D. Adams. This canvas is mounted in the Astro- 
nomical Room which is kept darkened but with the corona 
painting illuminated by indirect lighting. Those who have 
been privileged to see this painting have pronounced it a 
thing of rare beauty. The astronomers who saw the 1918 
eclipse and who have seen the picture look upon it as a marvel 
of perfection, true both as to form and color, a great work 
of art which has the added advantage of being scientifically 
accurate. In the same room of the great Museum are paint- 
ings of the 1923 and 192 5 coronas, also from the skilled hands 
of Mr. Butler, the three coronas forming a triptych with in- 
teresting contrasts. A reproduction of the 1923 painting is 
found facing page 204 and that of 1925 as the frontispiece 

The results from the expeditions of 1919 and 1922 which 
were devoted mainly to tests of the Einstein theory of rela- 
tivity will be found in Chapter XXIII. 



WHAT a splendid opportunity seemed to await the 
American astronomer at the eclipse of 1923! On 
September 10, the moon's shadow touched the 
earth's surface at sunrise in the Pacific off the coast of Japan. 
The shadow traversed the ocean at a speed well over a thou- 
sand miles per hour and appeared off the coast of southern 
California somewhat after noon. After crossing Lower Cali- 
fornia, Mexico and Yucatan, the eclipse ended at sunset in 
the Caribbean Sea north of British Guiana. 

It seemed very fortunate that the total eclipse track was 
to pass over a portion of California where one naturally 
expects during the summer months superb conditions of 
weather. Everyone, indeed, has heard of the boasted cli- 
mate of southern California, the " land of sunshine and flow- 
ers." In order to supplement the regular observations of 
the U. S. Weather Bureau, the Eclipse Committee of the 
American Astronomical Society for four years previous to 
1923 had had special observations made of the wind and 
weather during the first two weeks of September and at the 
hour of the eclipse. As the result of all the information avail- 
able it seemed that one of the best spots for an eclipse expedi- 
tion was San Diego, the only large city in the United States 
inside the path of totality. Within the memory of the " old- 
est inhabitant " there had not been a single cloudy day on 
the tenth of September. To make matters almost ideal, the 
total eclipse came at one o'clock when the danger from sea- 
fog was reduced to a minimum. A conservative estimate 
placed the chances of perfect conditions at least ninety per 
cent. For the first time in the history of science, the astrono- 
mer was able to insure his expedition from ill-luck of any 
kind, from clouds, from a sudden gust of wind that would 



shake the instruments or from any lack of adjustment that 
would endanger the final perfection of the photographs. Al- 
though eclipse insurance was offered to the expedition from 
the Leander McCormick Observatory, and at a very small 
premium, it was not accepted, for it seemed poor business to 
add to the expenses by insuring against the clouds that 
weather statistics showed to be so improbable. 

The author had made arrangements to observe his fifth 
total eclipse of the sun as a member of the party from the 
United States Naval Observatory in the event that the gov- 
ernment institution should send its own expedition. A spe- 
cial eclipse appropriation not having been passed by Congress 
and a Naval Observatory expedition thus being impossible, 
the necessary spectroscopic apparatus was loaned to the 
Leander McCormick Observatory. To add to his kindness, 
the superintendent of the Naval Observatory personally made 
arrangements with the Secretary of the Navy and the Chief 
of the Bureau of Equipment for the observers and equip- 
ment to be carried by naval vessel from Hampton Roads on 
the Atlantic to San Diego. 

On arrival in San Diego it was found that the U. S. Army 
was not to be outdone in generosity by the Navy. Before 
leaving the East, permission had been secured from the 
Chief of Coast Artillery for the University of Virginia party 
to locate within the military reservation of Fort Rosecrans. 
To put it mildly, the writer was amazed to have the com- 
manding officer of Fort Rosecrans extend an invitation, on 
behalf of the government, for the Virginia party to take up 
residence in officers' quarters in the Army post. Accord- 
ingly, six weeks before the eclipse, the McCormick Observa- 
tory party found themselves in a furnished house of ten 
rooms and two baths, the former home of a major. Meals 
were sent in from the company kitchen and Rosecrans 
boasted of the " best cook in the Army "! 

This army post is located on Point Loma, a peninsula lying 
between the Pacific Ocean to the west, and San Diego Bay 
on the east. On the top of the ridge which runs out to the old 
Spanish lighthouse is a famous drive, and from the Point it- 
self there is afforded one of the finest views in the world 


overlooking San Diego Bay and the city, North Island (with 
Rockwell Flying Field and Naval Air Station), Coronado 
Beach and Tent City; to the south lay the Coronado Islands, 
belonging to Mexico; to the east and southeast the moun- 
tains; to the west the blue water of the Pacific. 

San Diego has every reason to be proud of her city, located 
as it is in one of the finest spots on the globe and blessed with 
an agreeable climate that might well be the envy of any 
city in the world. We from Virginia, a state which has never 
been backward in painting a halo around everything con- 
nected with the Commonwealth, were much interested in the 
spirit of civic pride and aggressiveness, which at times may 
have even bordered upon boastfulness. Our six weeks' stay 
was filled with pleasant memories and we only wish that each 
and every total eclipse might be observed under the con- 
genial surroundings of Fort Rosecrans. 

Many are the problems that may be profitably attacked 
at the time of a total eclipse. One of the most important 
at the 1923 eclipse was the confirmation of the bending of 
the rays of light from a star as these rays pass close to the 
edge of the sun, as predicted by Einstein. We had no equip- 
ment for this research and we preferred to let others tackle 
this problem while we devoted ourselves to following up the 
line of attack carried out at four previous eclipses. It was 
therefore decided to have two eclipse stations, each equipped 
with a powerful spectrograph. One of these was located on 
Point Loma, the other at Lakeside, twenty-five miles inland, 
and near the edge of the shadow cast by the moon. The 
essential part of each spectrograph was a concave grating 
ruled by Rowland of Johns Hopkins. Each of the gratings 
used was made by ruling on a spherical concave mirror of 
speculum metal 15,000 lines to each inch. 

At North Island was the powerful battle squadron of the 
U. S. Naval Aircraft forces. Here was a chance to employ pho- 
tography from the air on any of the problems that could be 
solved by this method. It is manifestly impossible to use any 
but comparatively small cameras from an airplane and to give 
but very brief exposures. On account of the short exposures 
permitted no spectroscopic work could be attempted from the 


air, no investigation of the Einstein effect and no photog- 
raphy of the corona that demanded large focal scale. Air- 
plane photographs could not compete with those taken from 
a fixed installation on terra firma. In the event of clouds 
and the possibility of soaring above them in a machine, air- 
plane photographs might be taken, but there would be little 
of scientific value in photographing the corona on such a 
small scale. There seemed only one direction in which 
photography from the air could assist the astronomer, and 
that was in the attempt to find the position of the moon in 
the sky with greater accuracy. 

The program finally adopted consisted in attempting, from 
five separate stations along the northern edge and one at the 
southern edge on the shore of Mexico, to photograph from 
the air the ground underneath intersected by the edge of 
the shadow of the moon. For this purpose it was necessary 
to use the best of mapping cameras known to the photog- 
raphers and to choose special sites to photograph where the 
terrain would offer as great contrast as possible between a 
point just inside and one just outside the moon's shadow. In- 
cidentally it must be admitted that so little was known of the 
amount of light to be expected a few yards outside of the 
moon's shadow that there was some doubt as to the final suc- 
cess of the investigation. But there was nothing to do but 
to " try and see what happened." To supplement this work 
from the air, there were two parties of sailors, each of seventy- 
five, located at the expected edge of the moon's shadow. The 
sailors were put at intervals of ten feet at right angles to the 
edge of the shadow, and each was instructed to note whether 
he could see the corona or not. If he could see the corona 
during totality, he was presumably inside the penumbra of 
the moon's shadow. 

In addition to the pilots with mapping cameras, there were 
others with movie cameras stationed over Point Loma to pho- 
tograph the coming of the eclipse, each provided with a navy 
chronometer-watch to detect the accurate time, while an- 
other plane went aloft equipped with self-recording appara- 
tus to register temperature, humidity, etc., during the total 


During the erection of our apparatus it was interesting to 
watch the gradual installation of the gigantic equipment to 
be used on eclipse day near the tip of Point Loma by the as- 
tronomers from the Mt. Wilson Observatory. This great 
observatory, the best equipped and most famous in the whole 
world, is located on Mt. Wilson near Pasadena. Being there- 
fore only about 140 miles from the eclipse site, it was possible 
to transport all of their instruments by motor truck from 
observatory shops to eclipse camp. The problem of trans- 
porting the heavy apparatus was a very simple one for 
the Mt. Wilson staff, and one with which they are very 

To continue the magnificent work begun by Michelson in 
measuring the angular diameter of the star Betelgeuse, it was 
found to be necessary to separate the plane mirrors of the 
interferometer to the great distance of fifty feet. As this 
meant too large a span to be attached to the loo-inch reflector, 
a separate mounting was devised. This consisted in a bridge- 
like truss attached to a heavy polar axis driven by clock 
mechanism. The clock and central section of the mounting 
having been completed, the director of the Mt. Wilson Ob- 
servatory decided to utilize this interferometer structure on 
which to mount cameras, spectroscopes, photometers, etc., for 
the eclipse program. Here was a startling innovation in 
eclipse work, to put all of the instruments on one mounting, 
" the eggs all in one basket " but the scientific world has 
learned to have confidence in the judgment of the Mt. Wil- 
son astronomers. Their longest camera was thirty feet in 
focal length, used to photograph the Einstein effect. An- 
other camera half this length was for the same purpose. 
There were cameras to portray the beauties of the corona in 
various scales and in different colors of light, spectroscopes 
to ascertain the constitution of the corona, instruments to 
photograph and to measure visually the intensity of the light 
of the corona at various angles from the edge of the sun. The 
instruments (see page 208) numbering fifteen made prob- 
ably the most complete equipment that had ever been assem- 
bled for photographing a solar eclipse. The personnel in- 
cluded about thirty members of the Mt. Wilson staff while an 


auxiliary party of twenty computers and friends of the staff 
were prepared to watch and measure shadow bands, etc. 

To the astronomers gathered for the eclipse the weather 
seemed " unusual " (we understand this is the first word 
that a California baby learns at its mother's knee). At any 
rate the cloudless skies that we had been led to expect at 
1:00 P.M. were not always forthcoming. During the first 
two weeks of our stay at Point Loma, high fog at noon was 
the rule rather than the exception. Still it was a long time 
to the eclipse, and conditions would undoubtedly improve 
and they did. The next two weeks gave perfect skies, an 
absence of wind and altogether ideal conditions. Would 
this last, or would another cloudy spell come? We were 
optimists and believed implicitly what our friends the Cali- 
fornians told us. 

Saturday, September 8, was a cloudy day and the eclipse 
would have been under poor conditions. Sunday was even 
worse. What would Monday bring? Lieutenant Wyatt in 
charge of the weather observations at North Island made 
daily predictions of weather conditions to serve as a guide 
for airplane flights. He telephoned me twice on Sunday. His 
report was, " There has been excessively hot weather in Im- 
perial Valley for the past week. A secondary low over North- 
ern California makes the prospect of clear weather tomorrow 
very doubtful. If the back of the hot wave is broken there 
may be a change. This will be signalled by a brisk wind from 
the west. Unless this wind springs up in the night, I am 
afraid for you tomorrow. " 

At eight-thirty eclipse morning we were off on our trip to 
Lakeside. As we got inland from the Pacific, out of reach 
of the high fog, and where ordinarily we got out of the 
clouds, the conditions improved but little. At ten o'clock 
at Lakeside we found high cirrus and the conditions looked 
hopeless. But there were three good hours until the time 
of totality and much to be done in the final preparations! 
By twelve o'clock we heard the airplane aloft which was to 
observe over Lakeside and the pilot waved to us in friendly 
greeting. The clouds were not very heavy but there were 
no clear patches anywhere. We kept a stiff upper lip and 



refused to believe that after so much bragging California 
was going to get a black eye. 

Still the clouds were not too thick to prevent us watching 
the diminishing crescent of the sun as totality approached. 
Nature was hushed but the cocks were crowing lustily as if 
night were falling. At the very second when expected from 
the revised times sent out from the American Ephemeris, 
the sun was blotted out and a faint trace of corona appeared. 
But what a bitter disappointment ! We carried through our 
program, exposing eight plates in the seventy seconds of 
totality, knowing full well that the developed plates would 
show not the slightest trace of light. 

Well, there was no use crying over it (although one was 
tempted to). We had done our best and the fault was not 
ours. If misery likes company, it was evident that the clouds 
were general and not local. As quickly as a high-powered 
car could take us to San Diego we telephoned to Point Loma 
only to find as we expected, that conditions there were even 
worse and practically not a trace had been seen of the corona. 
Radio soon told us that the large assemblage of seventy as- 
tronomers on Catalina Island, including the Yerkes * and 
Harvard parties, Plaskett, Stebbins, Fox, Wilson and others, 
had suffered a like fate. Even the usual luck of the Lick 
party had deserted them, for conditions where Dr. Campbell 
and the Lick expedition were located at Ensenada 2 were 
about as bad as could possibly be. 

And this was the record for " sunny California "! Not a 
single expedition greeted with good conditions, and the whole 
scientific work a dismal failure! There was nothing left to 
do but pack up and go home and then begin to get ready 
for the next eclipse. 

The weather conditions in Mexico did not give promise 
of the superb conditions that had been expected in southern 
California, so it was all the more gratifying to find the good 
luck experienced by the astronomers in old Mexico. Four 
separate expeditions were wholly or partially successful. The 
Mexican government financed two expeditions, one from the 

1 Popular Astronomy, 32, 205, 1924. 
- Publications A. S. P., 35, 275, 1923. 


Mexican National Observatory of Tacubaya under the di- 
rection of Gallo, the other L consisting of Schorr, Luden- 
dorff and Kolhschiitter from Germany. The Steward Ob- 
servatory of Arizona had an expedition 2 in the province of 
Sonora and the Sproul party 3 was located at Yerbanis. Di- 
rector J. A. Miller had the capable assistance of Heber D. 
Curtis, director of the Allegheny Observatory. 

The Sproul expedition had a varied program of work, but 
mention will be made here of two items only, photographs 
with Einstein cameras of fifteen feet focus and photographs 
on a larger scale of sixty-five feet focus, the cameras in each 
case being pointed directly at the sun. The photographs for 
the Einstein effect, one of which is reproduced facing page 
212, showed the corona in splendid detail. The star images 
however were not of the best definition, and as the photo- 
graphs were made through thin haze, it was decided by Miller 
that he would not measure the plates to obtain the deflec- 
tions of the star images. 

After the ill luck experienced by the American astronomers 
in 1923 where conditions beforehand had seemed so promis- 
ing, it seemed almost foolhardy to prepare for an eclipse 
which was to take place at nine o'clock on a winter's morning 
on the Atlantic coast in the United States with the sun at best 
only eighteen degrees above the horizon. Would snow cover 
up the astronomers as had happened on January 14, 1907, 
the preceding eclipse in this cycle of the Saros? The most 
optimistic placed the chances of good weather about fifty 
per cent. With such poor prospects it was natural that none 
but American astronomers would plan to observe the eclipse. 
In fact, the chances of success seemed so remote that the 
Lick Observatory, with its splendid record of carefully 
planned observations secured at many eclipses, decided 
against sending an expedition across the continent. After 
attempting a very large program at the 1923 eclipse, the 
Mount Wilson Observatory contented itself in 1925 with at- 
tacking about one-fifth of the problems planned for the 
previous eclipse. 

1 Sitz. Preuss. Akad., 83, 1925. 

2 Publications A. S. P., 36, 170, 1924. 

3 Astroph. Jour., 61, 73, 1925; Sproul Obs. Publ., No. 7, 


What a surprise the weather again had in store for us on 
January 24, 1925! The day before the eclipse was one of 
gorgeously perfect blue skies. Would the morrow provide 
equally good skies? The clear skies continued throughout 
the night but it clouded over completely at six in the morn- 
ing and totality occurred three hours later. The largest 
group of astronomers was assembled at Middletown, Con- 
necticut, at the Van Vleck Observatory, and here were parties 
representing the Mount Wilson Observatory, Leander Mc- 
Cormick Observatory, Harvard University, Universities of 
Wisconsin and Illinois, United States Bureau of Standards, 
and many others. What a dejected crowd we were at eight 
o'clock when we had gathered at the Van Vleck Observatory 
to observe first contact, the beginning of the eclipse. There 
was nothing but clouds everywhere! A quarter of an hour 
later a ray of hope appeared; there was a blue streak of sky 
low down in the northwest and the clouds were coming 
from that quarter. Would it clear off in time? Luck was 
with us. Five minutes before totality a cloud, very thin and 
very fleecy, hung over the sun. It was not thick enough to 
do much damage and it was moving slowly. We hoped it, 
too, would go. When the timers called out " two minutes," 
the cloud was almost gone. Now it was beginning to get 
quite dark, a weird and unnatural pall coming over the land- 
scape. The observers outside noticed shadow bands flicker- 
ing over the snow. At one minute before totality, with the 
thin crescent of the sun growing very small, the atmospheric 
conditions seemed perfect, the thin cloud had gone! 

How fortunate we were that the weather defied all the 
laws of averages on January 24, 1925! With clouds hover- 
ing everywhere over New York and New England, the sur- 
prising fact was that clear skies greeted most of the as- 
tronomical expeditions. It was cloudy throughout Michigan 
and Ontario, cloudy in Buffalo, but clear at Ithaca, Pough- 
keepsie, Middletown, New Haven, Nantucket and New York. 
It is estimated that ten million people in New York State and 
New England were given an opportunity of witnessing the 
gorgeous spectacle. The great newspapers of New York City 
asserted that no single event in the past decade has aroused 


such widespread enthusiasm. As totality lasted for the brief 
time of two minutes or less, and as the scientific investiga- 
tions were nearly all crowded within the period of the total 
phase of the eclipse, it is probable that no single event in the 
history of man has had so many words, per minute of the dura- 
tion of the event, written about it as has the 1925 eclipse. 
As a spectacle this eclipse suffered from taking place so early 
in the morning. If the darkening had come on during the 
middle of the day with the sun high up in the sky, the psy- 
chological effect would have been greater. The shape of the 
corona corresponded closely with that expected from the 
condition of the sun near spot minimum. To the right of 
the vertical, however, there was a long pointed shaft of light 
stretching up more than a degree. The remarkable feature 
was the total lack of rosy color visible to the naked eye, no 
prominences being readily visible, and as a result the corona 
lacked color as if, possibly, to reflect the feelings of the ob- 
servers who everywhere worked with the thermometer below 
the zero mark of the Fahrenheit scale. 

Several unusual features concerning this eclipse are worthy 
of note. The eclipse track crossed over the Van Vleck Ob- 
servatory with its visual refractor of 20 inches aperture. 
Here was an opportunity of photographing the corona with a 
large telescope and with yellow light. Miller of the Sproul 
Observatory used the photographic lens of 63 -feet focal 
length, mounted as in 1918 and 1923, pointed directly at the 
sun. Naturally there were hosts of cameras of smaller focal 
length. On account of the high possibility of clouds, the 
United States Navy employed the dirigible Los Angeles which 
flew out to sea and carried a battery of cameras with which 
to photograph the corona. Even in a gigantic dirigible the 
platform carrying the cameras is not very steady and it is 
very difficult to keep the cameras pointed at the sun. Natu- 
rally it was possible to utilize the cameras only of short focal 
length and to give brief exposures only. If it had been 
known beforehand that the skies were to be clear at eclipse 
time, the Los Angeles would not have taken part in the 
eclipse program. Its photographs showed the general fea- 
tures of the corona but they had little of scientific value com- 

Sproul Observatory photograph with 15 -foot camera. 



4 3 
- O 





pared with those taken by cameras with more stable founda- 
tions and greater focal lengths. Many airplanes were utilized 
for the same mission. 

Many attempts were made to measure the intensity of the 
corona, among which should be noted the work of Nicholson 
and Pettit, by Stebbins and King, and by Coblentz and Stet- 
son, all three parties gathered in Middletown; and also by 
Parkhurst, located farther west in Ithaca. Curtis at New 
Haven, Anderson and Mitchell, both in Middletown, photo- 
graphed the flash spectrum, the first named paying special 
attention to the red end and reaching to wave-length 8800 A 
on the photographs. The results obtained by these various 
observers will be duly recorded on the appropriate pages 
later in this book. 

The developed photographs all show the effect of the low 
altitude of the sun and the consequent poor seeing. As 
eclipses are of such rare occurrence it is fortunate that coronal 
details do not require the finest conditions of seeing for their 
precise portrayal. The details of the coronal streamers are 
not inherently sharp and clear-cut in their nature, and conse- 
quently a little blurring of detail caused by poor atmospheric 
conditions has little deleterious effect on the photographs. 
The case is, however, very different with the prominences. 
By nature these phenomena have more definite outlines. The 
best photographs with which to test the quality of the seeing 
at the 1925 eclipse were unquestionably those taken with 
the 20-inch refractor of the Van Vleck Observatory. This 
telescope is regularly employed in the determination of stellar 
parallaxes bv photography. The focus consequently has 
been thoroughly well determined. The plates taken by this 
telescope (which has the largest aperture ever used to pho- 
tograph an eclipse) were a great disappointment, due mainly 
to the lack of sharp detail shown in the inner corona. 
The fault did not lie in the telescope or in any lack of care- 
ful adjustment, but was to be found in causes beyond the 
control of the observer, namely, the poor definition resulting 
from bad seeing. The same qualities of poor seeing are found 
in the spectral images of these prominences in the flash spec- 
trum taken without slit. 


At the eclipse of 1923, the preparations made to photo- 
graph from airplanes, so as to fix with greater exactness the 
place of the moon in the sky, came to naught because of 
clouds. In 1925, on account of the southern edge of the 
moon's shadow passing over New York City, an excellent 
opportunity was at hand to secure many observers to deter- 
mine the exact edge of the shadow. 

Calculations based on the American Ephemeris had fore- 
told that the edge of the moon's path would cut Riverside 
Drive (which runs north and south along the Hudson River) 
somewhere between 83d Street and noth Street, with a 
total uncertainty of approximately one mile. To make cer- 
tain that the astronomers were not mistaken, observers were 
located at each intersection of city blocks all the way from 
72d Street to i3Sth Street, usually on the tops of apartment 
houses, so that a better view might be obtained. Sixty-nine 
men were employed and each was furnished with a piece of 
darkened glass and was instructed to look at the sun at the 
time of totality in order to see whether the corona was visible 
or whether there was a thin edge of the sun left shining. All 
of the observers were instructed to report to a central office 
immediately after the eclipse. Only one of the total of sixty- 
nine was in doubt as to what he saw, and the sixty-eight gave 
a clear-cut verdict. The observer at 240 Riverside Drive 
had seen the total eclipse while a man located at 230 River- 
side Drive had seen a small sliver of the sun exposed, indi- 
cating that it was a partial eclipse. The distance between 
these two men was about two hundred and twenty-five feet, 
which included the width of 9 6th Street. The edge of the 
shadow of totality was, therefore, pinned down to within 
two hundred and twenty-five feet on the west edge of Man- 
hattan Island. Other observers along the East River were 
successful in making similar observations with the result 
that the moon's path across New York City is accurately 
known. It is indeed surprising to find such unanimity of 
opinion among untrained observers who were all witnessing 
their first eclipse. Apparently it must be very easy to make 
up one's mind as to whether the corona is or is not visible. 
Similar attempts have been made at former eclipses but have 


always met with failure. None of the observers saw the 
edge of the moon's shadow as it lay upon the ground in spite 
of the excellent opportunity afforded on account of the ground 
being completely covered with snow. Evidently the edge 
of the shadow is not sharply defined but the light tapers off 

The keen interest of astronomers in total eclipses of the 
sun again manifested itself in the Dutch East Indies on 
January 14, 1926. No less than nine expeditions from 
Europe and America were gathered for the purpose of work- 
ing ardently and enthusiastically for the brief period of four 
minutes. Three expeditions, numbering twenty people, trav- 
eled from the United States halfway round the globe in the 
hope of having clear skies during the few precious minutes 
of totality. 

In recounting the history of eclipses, repeated attention 
has been called to the many failures of eclipse expeditions 
due to the fickle weather. This has been especially true in 
the tropics. Where the writer was located in Sumatra for 
the 1901 eclipse the annual rainfall was the goodly amount 
of 1 86 inches or an average of half an inch per day. 

Unfortunately, at the 1926 eclipse the weather lived up to 
its poor reputation. Clouds were the rule rather than the 
exception. The British party at Benkoelen on the west 
coast of Sumatra had the best luck. Their program 1 was 
almost exclusively spectroscopic. Excellent spectra were 
obtained which will be described in a subsequent chapter. 

Although the party from Swarthmore 2 was located not far 
from the British in Benkoelen, they did not fare so well with 
the weather. Successful photographs were secured with the 
large camera of 63 -feet focal length pointed directly at the 
sun, with twin cameras of i5-feet focus for testing the Ein- 
stein deflection, and with smaller cameras. Curtis again used 
the short focus concave grating for the flash spectrum but on 
account of the haze and the deterioration of the plates 
stained with dicyanine no spectra were obtained. 

The Harvard expedition, consisting of Stetson, Coblentz 

1 Memoirs R. A. 5., 64, 105, 1927. 

2 Popular Astronomy, 34, 349, 1926. 


and Arnold, measured the intensity of coronal radiation visu- 
ally and photographically. Their results seemed to show 
that the total coronal light was 40 per cent greater in 1926 
than in 1925. However, their work was done at the site 
of the Swarthmore party and it is possible that the haze may 
have affected their results. (See Chapter XXI.) 

On the other side of the mountain range from Benkoelen 
was the expedition from the U. S. Naval Observatory includ- 
ing Anderson of Mount Wilson with a 21 -foot concave grat- 
ing for the flash spectrum. The eclipse was practically lost 
through clouds covering the sun for nearly the whole of 
totality. Freundlich of Potsdam located at Benkoelen on 
account of haze secured few photographs of value. In East 
Africa, Horn d'Arturo l at the head of a completely equipped 
Italian expedition was more successful with the weather. 
The photographs of the corona taken by him were compared 
with those obtained in Sumatra. Many changes were noted 
giving evidence of motions in the coronal structure, particu- 
larly in the domes which are such interesting phenomena at 
each eclipse. 

On June 29, 1927, the eclipse track passed over England, 
Norway, Northern Sweden, the Arctic Ocean and North- 
eastern Siberia. Unfortunately the eclipse was a very brief 
one. At Liverpool, totality took place at 5:24 A.M. and 
lasted only twenty-three seconds. At Fagernes, Norway, 
totality occurred at 5:34 (Greenwich time) and had a dura- 
tion of 34 seconds. To make matters worse, the probability 
of clear weather was not very great. In England, at the 
early hour of the morning that the eclipse took place, the 
most optimistic estimated that the chance of clear weather 
was no more than one out of three. In Norway, away from 
the coast and farther along the eclipse track in Lapland, the 
chances of clear weather appeared to be greater but at best 
were no more than an even chance. It seemed almost fool- 
hardy to attempt to observe this eclipse with such a short 
duration of totality and with the chances of clear weather 
reduced to the minimum. Still nothing venture, nothing 
erain! It was certain beforehand that no great astronomical 

1 Publ. d. Oss. Astron. di Bologna, i, 227, 1926. 


discoveries would be made from coronal investigations with 
this short duration. Still it was very important that a record 
should be made of the eclipse. For photographing the 
flash spectrum the short duration of totality was no great 

The popular enthusiasm in England was unbounded. Eng- 
land had not witnessed a total solar eclipse for more than 
two centuries (see page 58) and the populace was deter- 
mined to attempt to see the wonderful phenomenon no mat- 
ter how dismal were the chances of success. Nearly all of 
the British astronomers were in the eclipse track in England 
with well-planned expeditions. A large and well-equipped 
party from the Solar Physics Observatory of Cambridge, 
however, forsook England for the greater promise of clear 
skies in Norway and set up their equipment at Aal under 
the direction of H. F. Newall. The only large expedition 
from the United States was located in Norway at Fagernes, 
fifty miles across country from the Cambridge party. The 
McCormick-Chaloner party from the University of Virginia 
had a very complete equipment to photograph the flash spec- 
trum with three concave gratings, each used without slit in 
the attempt to secure the flash r spectrum from the extreme 
ultra-violet to the far red, the plates for the region of long 
wave lengths being stained with neocyanine. Stetson was 
a member of the party at Fagernes with the instrumental 
equipment he had used at the eclipses of 1925 and 1926 for 
measuring the radiation of the corona. The astronomers 
from Oslo were also at Fagernes. Rosseland had a 21 -foot 
concave grating for photographing the flash spectrum, Lohse 
attempted to observe the times of contacts visually and by 
photography, Stormer had a number of small cameras for 
photographing the corona. A scientific party of about forty, 
of ten different nationalities, were assembled at Fagernes. 
Other parties were located farther along the eclipse track, 
mainly at Ringebu in Norway, and at Gallivare and Jokk- 
mokk in Lapland. 

Both in England and in Scandinavia the weather before 
the eclipse was much worse than had been expected. At 
Fagernes, instead of being clear for half the time, it had 


been cloudy for fourteen days before eclipse day and 
rain had fallen on nine of these days. Decided optimism 
was required in order to expect a sudden break in the weather 
conditions. In England in addition to the cloudy weather, 
there were very high winds that threatened to carry away 
the temporary installations. At this date of the year, June 
29, and at the high northern latitudes, the adjustment of the 
eclipse instruments had to be accomplished without the 
stars. At Fagernes (latitude 61) it did not get dark at 
night, while Jokkmokk was in the land of the midnight sun. 
Eclipse day dawned everywhere with the weather a con- 
tinuation of the cloudy conditions of the previous fortnight. 
Apparently only a miracle could save the situation and bring 
clear skies. The miracle actually did take place in England 
at Giggleswick, where was located the party from the Green- 
wich Observatory under the direction of the Astronomer 
Royal, 1 Sir F. W. Dyson. Here the sky was cloudy through- 
out the whole of eclipse day. However a hole appeared in 
the cloudy sky and lasted for a space of two minutes only. 
The eclipsed sun appeared in this blue patch of sky and suc- 
cessful photographs were secured. The photographs show 
exquisite detail in the inner corona but the outer corona was 
lost through thin haze. At Jokkmokk where was located 
the expedition from Hamburg under the direction of Schorr, 
a similar miracle took place, the sun at totality being sur- 
rounded on all sides by clouds. Schorr was more successful 
than Dyson, the sky being without haze. The photographs 
taken at Giggleswick and Jokkmokk exhibit many brilliant 
prominences with the hooded forms overtopping the promi- 
nences showing as conspicuous features. A comparison of the 
two photographs shows well-marked changes taking place 
during the time elapsed. Schorr's photographs show the 
corona to be of the circular type that was expected at the time 
near maximum of sun-spots. The brilliant prominences and 
extended corona must have made a gorgeous spectacle for 
those few fortunate souls who were lucky enough to behold 
the eclipsed sun under good conditions. Elsewhere along the 
track thin haze or heavy clouds interfered with the work or 

1 Monthly Notices R. A. S., 87, 657, 1927. 


made the efforts of the astronomers of no avail. At Gallivare x 
in Sweden, there were several expeditions, the party from 
Upsala securing photographs of the corona and of the flash 
spectrum. These observations were secured through thin 
clouds. The flash spectrum was photographed with success 
by Pannekoek and Minnaert," by Vegard, and by Baade. 

At the 1925 eclipse the attempts which were made to 
photograph the phenomenon in color met with partial suc- 
cess. At the 1927 eclipse a color film was taken at Giggles- 
wick making a brief exposure every second and lengthening 
these to about 0.5 second as totality approached. The results 
are described by the Astronomer Royal. 3 At Giggleswick the 
duration of totality was 23 seconds. For 27 of the exposures 
by the color film it may be said that the eclipse was total, and 
for an additional two seconds at each end Baily beads are 
shown. The corona and prominences can be traced on the 
limb of the sun opposite to the crescent for a duration of 30 
seconds both before and after totality. The corona shows on 
the film to be somewhat bluish in color during totality. 
Outside of totality this bluish color appears rather reddish, 
the difference in color being due to the process employed. 
The chromosphere appears quite red in color while the promi- 
nences vary from a deep to a whitish red. 

Before totality the corona appeared in some places bluish 
and in others reddish. When totality commenced there was 
a continuous reddish arc of the chromosphere on the east 
limb, extending for about 130 with red prominences. The 
chromosphere was visible on the film for 9 seconds after the 
beginning of totality and reappeared again 6 seconds before 
the end of totality. The colors on the film do not agree with 
those seen by the eye. 

On account of the short duration of totality the bright 
inner corona was left exposed beyond the dark limb of the 
moon for a longer space of time than is ordinarily found at 
a total eclipse of the sun. The result of this has been that 
the corona was seen and photographed for a much longer 
stretch of time outside of totality than is usually the case. 

1 Kienle, Die Himmelswelt, 37, 225, 344, 1927. 

2 Verhand. der Kon. Akad. van Wetenschappen te Amst'dm., 13, No. 5, 1928. 

3 Dyson, Monthly Notices R. A.S., 88, 142, 1927. 


References to other features of the 1927 eclipse will be 
found at the appropriate places in the following pages. 

The next eclipse, May 9, 1929, was visible in northern 
Sumatra, the Malay States, Siam and the Philippines. 
Many expeditions were attracted to the eclipse track in spite 
of the great distances from home that must be traveled and 
the uncertainty of the tropical weather. Professor J. A. Mil- 
ler, a veteran of many eclipse expeditions, tells 1 of the gener- 
osity of the Dutch in caring for the work of the astronomers. 
All of the heavy equipment was carried free of charge by the 
Dutch steamship lines from the United States halfway round 
the world to Sumatra and return, while the official personnel 
of the expedition was given passage at half the regular rates. 
Unfortunately the Dutch could not control the weather 
in Takengon in northern Sumatra where was located the 
Swarthmore party and also the Potsdam expedition under 
Freundlich. Miller describes the weather as follows: "We 
had not seen a clear sun for days preceding the eclipse. It 
was at that season of the year known as the change of the 
Northeast to the Southwest Monsoon. The Southwest Mon- 
soon was to bring clear weather, and it did, five days after 
the eclipse. 

"The eclipse occurred at 12:47 o'clock. In the morning 
the sky was entirely overcast. About nine o'clock a round 
patch about 20 in diameter appeared in the southeast sky. 
It persisted, and seemed to drift slowly toward the com- 
pletely overcast sun. At eleven o'clock one could tell where 
the sun was in the sky. At the time of first contact the sun 
was invisible, but the blue circular patch still drifted slowly 
on. Sometimes during the partial phase one could tell by 
using field glasses that the sun was eclipsed. Ten minutes 
before second contact we took our places at the instruments, 
having decided to make all exposures regardless of the 
clouds. At that time one could not tell with the unaided eye 
that the sun was in partial eclipse. Five minutes before 
totality it was considerably brighter, and at totality the sun 
was in the center of the circular blue patch, which to the 
naked eye seemed perfectly clear. It remained clear during 

1 Popular Astronomy, 37, 495, 1929. 


the entire period of totality, then the clear patch moved on, 
and half an hour after totality the sun was again obscured by 

The Swarthmore party secured successful photographs 
with the 63-foot tower camera, with a pair of 1 5-foot Ein- 
stein cameras and with smaller instruments. The German 
party secured excellent photographs for the Einstein effect 
and also spectra of the corona in fine definition. To obtain 
check plates the party stayed behind at the eclipse site. 
Observers elsewhere had varied success. 

Two British expeditions were located at Alor Star in 
Kedah and at Pattani in Siam. The former photographed 
through haze, the latter was blotted out by clouds. Stetson, 
Arnold and Johnson at Alor Star with an illuminometer 
found at mid-totality a radiation of 0.15 foot candles. The 
U. S. Naval Observatory expedition was at Iloilo in the 
Philippines. Good photographs of the corona were secured 
through light clouds. In Cebu the Hamburg party had thin 
clouds; Dutch, French and Japanese expeditions were par- 
tially successful. 

The annular-total-annular eclipse of April 28, 1930, pre- 
sented special problems. In consequence of the usual devia- 
tion of the moon from its predicted place, would it be 
possible for an expedition to succeed in locating itself inside 
the path of totality scarcely more than a half mile in width? 
Evidently, at an eclipse of this character the most fruitful 
field of investigation would be the flash spectrum. 

Heavy rain and clouds along the track in California and 
Nevada greatly interfered with the observations. With their 
customary good fortune, the Lick observers had a clear sky 
at the time of totality. Two prism spectrographs were em- 
ployed each with moving plate according to the method 
developed by Campbell. The photographs by both instru- 
ments ] recorded the lines of intermediate and high levels but 
the timing was too late to obtain the reversal of the lines of 
lowest level. The expedition from Mount Wilson was ham- 
pered bv clouds. Photographs of the flash spectrum were 
obtained by two concave grating spectrographs, one of 

i Publ. A. S. P., 42* 131, 1930. 


io-foot radius used with moving plate and the other of 
2 1 -foot radius with minified images of the crescents. The 
latter arrangement designed by Anderson was used at the 
1926 and 1927 eclipses. The detailed results of the spectra 
from this brief eclipse will be watched with interest. 

The path of the recent total eclipse started at sunrise on 
October 22, 1930, in the North Pacific and ended at sunset 
off the coast of Patagonia on October 2 1 . The observation 
of the eclipse was dramatic and spectacular in all of its 
details. The only available site was a small, isolated, vol- 
canic island not far from the International Date Line and 
at 15 south latitude. The U.S. Naval Observatory secured 
a small appropriation from Congress for an expedition. As 
a result of their commendable policy, frequently put into 
effect in the past, astronomers with eclipse experience were 
invited to become guests of the expedition. The equipment 
consisted of 115 cases of scientific instruments and supplies, 
60 tons of stores and 11,000 feet of lumber. A naval vessel, 
U.S.S. Tanager, was detailed to assist in the work. The 
members of the party were on the island two months before 
the eclipse. 

The island of Niuafoou in the Tonga group is familiarly 
known as " Tin-Can Island. " The island being both small 
and volcanic there are no bays nor landing places where a 
boat could run alongside. Once each month there is great 
excitement on the island; the mail steamer arrives! Unfor- 
tunately, the only sure method whereby the mail can be 
brought ashore is to seal it up in a tin can and lower it over 
the ship's side; then one white trader and two natives swim 
out for the tin can at the same time carrying the out- 
going mail. But the island is in the path of the trade 
winds, and if the water is too rough to permit the swim- 
mers to go out, the mail steamer does not stop, and the island 
gets along without news from the outside world until an- 
other month rolls by. 

The island is shaped like a gigantic signet ring. It has a 
diameter of five miles, with an interior lake three miles in 
diameter whose surface is about seventy-five feet above sea 
level. The inhabitants, uoo in number, are natives of the 


Polynesian race. When we were on Niuafoou the white male 
population consisted of three persons, two white traders and 
a Catholic priest. 

The island boasts of the largest cocoanuts in the Pacific 
Ocean. Copra or the dried ripe cocoanut is the only article 
of commerce. When a boy reaches the age of eighteen he 
is given eight acres of land by the Government; when he 
marries he is given a plot of ground in the village for his 
home. The meat of the young green cocoanuts provides 
food for man, dog, chicken and pig, and the water is used 
for drinking and cooking, and also for the preparation of the 
ceremonial drink of kava. 

As the cocoanuts grow readily without any attention, life 
for the South Sea Islander is very simple. The government 
taxes must be paid before the end of September and hence 
in the early days of the month there is great activity in 
gathering the ripe cocoanuts, drying the meat, and then sell- 
ing the supply to one of the two trading companies. Then 
comes the tramp steamer from the outside world to load the 
copra to take it to the United States for making food sup- 
plies and soap. When the taxes are once paid and a receipt 
obtained, the native may take it easy for the balance of the 
year. In contrast to life under the benefits of white civiliza- 
tion, the natives lead a care-free existence. No wonder that 
Stevenson and others have written such fascinating tales of 
life in the South Seas! We enjoyed greatly the friendly 
intercourse on our brief sojourn on Niuafoou. At four 
different times we were the honored guests at a grand ban- 
quet of native delicacies followed by a dance when we were 
given the privilege of dancing, American fashion, with the 
native belles. The daughter of the head man of the village 
where I had been adopted as " friend," rejoiced in the eupho- 
nious name of Vaicima which translated into English 
means " cement water-tank." 

The erection of the big cameras was no easy task. There 
was the 63-foot tower telescope, the camera of 65-foot focus 
placed horizontally with coelostat, a pair of Einstein cameras 
of 1 5 -foot focus on a separate equatorial mounting, together 
with smaller cameras. All of this was under the direct 


supervision of Professor Ross W. Marriott of Swarthmore. 
I had two concave grating spectrographs used at former 
eclipses. Fortunately, Commander Keppler and Dr. Kellers 
of the U. S. Navy had been at the eclipses of 1926 and 1929, 
and eclipse problems were to them no novelty. Two petty 
officers and eleven enlisted men from the Navy helped in the 
erection of the apparatus. A month before the eclipse the 
Americans were joined by a party of seven from New Zealand 
under the direction of my old friend, Dr. C. E. Adams. 

Recent eclipse expeditions to the tropics, in 1926 and 
1929, had been treated rather badly by the weather. What 
was in store for us? Weather records kept by us during the 
eight weeks preceding the eclipse had shown that only one 
day out of three was clear at the eclipse hour. On eclipse 
day, October 21 by the American date and one day later 
according to local reckoning, conditions before the great 
event were even worse than usual. Totality was to occur at 
8:51 in the morning. Two hours before this a light rain was 
falling; first contact was observed through clouds. The 
clouds soon after began to get thinner and fifteen minutes 
before totality they had entirely cleared away. The total 
phase was observed under nearly perfect conditions. Pos- 
sibly there was a slight haze but this in no way seemed to 
affect the photographs. One half hour after totality, clouds 
again began to gather! On the night following the eclipse, 
Marriott started on the development of the coronal photo- 
graphs. On the following night the dark room was given 
over to my exclusive use. As soon as it had cooled somewhat 
I started, and with the help of ice brought for the purpose, 
I was able to keep developing, fixing and washing baths at a 
fairly satisfactory temperature. The washing of the plates 
was difficult on account of the scarcity of water. On the 
island the only available fresh water came from rains. Un- 
fortunately there had been practically no rain during our 
residence on Niuafoou and all of the cisterns in the vicinity 
of the eclipse camp were dry. Fresh water had therefore 
to be brought by the Tanager in eight-gallon breakers and it 
was necessary to conserve this in every possible way. That 
used for washing the photographic plates was saved and was 



g "H 






utilized again for personal use and for dish washing. Lava 
dust permeated everywhere and it was difficult to dry the 
plates, especially when large in size, without their being 
spoiled by dust particles. 

When all of the photographic plates, direct corona and 
spectrographic, were developed, the eclipse camp was a 
cheery place. We had been fighting against tremendous odds 
but we had won out. I have never seen more exquisite 
detail than is found on the coronal photographs taken under 
the direction of Marriott. My own spectra left little to be 
desired, with good focus from wave-length 3200 in the violet 
to 7800 in the red and with the region from 4650 to 6800 in 
duplicate. Measures of coronal radiation and observations 
of shadow bands will be discussed in later pages. 

I had gone to Tin-Can Island, and then around the world 
on the return journey, to work for ninety-three seconds of 
time. On August 31, 1932, I shall observe my ninth total 
solar eclipse at a location in Magog, Quebec compara- 
tively speaking, almost at home. There also will be my 
friend Stratton at the head of an expedition from Cambridge, 
England, and Minnaert from Utrecht, Holland. Farther 
north along the track, at Parent in northern Quebec, will be 
an expedition from the Greenwich observatory. In northern 
New England there will be expeditions from the two Great 
California observatories, Mount Wilson and Lick, and also 
parties from Swarthmore, Van Vleck and other American in- 
stitutions. The eclipse will be total for about one hundred 
seconds of time and the path will be approximately 100 miles 
in width. The chances of clear weather appear to be about 

By referring to page 55 it will be seen that if I am ever 
to observe a tenth total eclipse of the sun it will be necessary 
for me to travel many thousands of miles from my home in 



DURING the past two and a half centuries, the as- 
tronomer and the mathematician have worked in 
close cooperation investigating the distances and 
motions of the bodies forming the sidereal universe. Under 
the magic wand of their combined labors, the complexity of 
the Ptolemaic system has given way to great simplicity and 
beautiful order, revealing motions obeying the inverse square 
law of Newton, or the very slight modification of this law 
of gravitation demanded by the theory of relativity. Dur- 
ing the progress of these investigations, our conceptions of 
distances and dimensions have been gradually modified, so 
that the sidereal universe appears to be vastly greater than 
was formerly thought. Our own Galaxy probably has a radius 
exceeding icr 1 meters or one hundred thousand light years, 
while beyond extend other universes at distances of hundreds 
of millions of light years. 

While the astronomer has thus been reaching out to 
greater and yet still greater distances in the direction to- 
ward the infinite, the physicist and the chemist, on the other 
hand, have found solar and planetary systems of nearly 
infinitesimal dimensions within the realm of the chemical 
atom. The radius of the electron we seem to think we know 
is equal to 1.9 x io- ir> meters. The astronomer has con- 
tributed much information concerning the atom because the 
celestial laboratories of sun and distant stars provide high 
temperatures and minute pressures transcending any avail- 
able in the best-equipped terrestrial laboratories. 

The amazing development of our conception of the atom 
has come within the past three and a half decades, the be- 
ginning taking place with the study of streams of negative 
corpuscles or electrons. On its discovery, radium seemed 



superficially to exhibit a contradiction of the laws of con- 
servation of energy. Heat and light were spontaneously 
emitted without any apparent changes in the radium, and 
thus a continuous supply of energy seemed evolved which 
set at naught that fundamental law of physics. But it was 
soon found that with the giving off of energy in radiation, 
the radium itself did utterly change and here the philoso- 
pher's stone seemed at last to have been discovered, for 
it was found that one chemical element actually changed 
into another. 

And to think that after all the whole science of radio- 
activity was more or less the result of a happy accident! 
The year following the discovery of X-rays in 1895 by 
Rontgen, Becquerel of Paris wished to test the phosphores- 
cent action of certain substances by wrapping a photo- 
graphic plate in black paper, and placing on it the substance 
to be examined, which was then exposed to sunlight. By 
good fortune a preparation of uranium was chosen and the 
photographic plate was darkened. The Becquerel rays were 
thus discovered, and it was soon found that, like the X-rays, 
these rays penetrate substances impervious to light, even 
passing through thin plates of metal. The experiments were 
always made by placing the phosphorescent substance in the 
sunlight on top of the black paper enclosing the photo- 
graphic plate. But one day the sun was clouded, and the 
plate and the phosphorescent substance were placed away 
in a desk and were left there for several weeks. Becquerel 
for some reason developed the plate, and was surprised to 
find the plate darkened as before, thereby showing that 
probably neither sunlight nor phosphorescence had any- 
thing to do with the action on the photographic plate. Thus 
was born, in 1896, the new science of radioactivity! 

Besides the effect on the photographic plate, radioactive 
substances manifest themselves in three different manners: 
first, by exciting phosphorescence and fluorescence; second, 
by causing the air near them to become conductors of elec- 
tricity; but most startling of all, by the continuous genera- 
tion of light and heat. 

Mme. Curie recognized that radioactivity was a property 


of the atom and starting with this in view she found that 
the residues from the mine at Joachimstal, Austria, were 
three to five times more radioactive than uranium. From 
this residue she separated out a new substance, far more 
active than uranium which she called polonium, in honor of 
the place of her birth. Later she discovered radium. This 
appears to be an element with atomic weight 226, and it is 
found in excessively minute quantities, there being only one 
part in five million in the best pitchblende. In 1899, 
Rutherford showed that the radiation from uranium was 
complex, consisting of (i) the a rays, which are absorbed 
by a sheet of paper or a few centimeters of air (2) of a 
hundred-fold more penetrating (5 rays, capable of passing 
through several millimeters of aluminium, and (3) of still 
more penetrating y rays, capable of passing through quite 
a thickness of iron and lead. The $ rays are deflected by 
a magnetic field. Becquerel and Kaufmann showed that 
the (3 rays were negatively charged particles projected with 
a velocity approaching that of light. The very penetrating 
y rays are not deflected in a magnetic or electric field, and 
are probably closely connected with X-rays. 

Though the a rays are the least penetrating, they are 
much the most important of the three types of radiation. 
They are deflected much less by a powerful magnetic field 
than the $ rays, and in the opposite direction, showing that 
the a rays consist of a stream of positively charged particles. 
Alpha rays, therefore, will affect a gold leaf electroscope, 
and this old instrument gives one of the most sensitive 
methods of measuring the amount of radiation. In fact, 
Rutherford has shown that it is not difficult to measure with 
certainty the presence of radium in a body which contains 
as small a quantity as io- n grams of radium! 

The maximum velocity of the a rays is 12,000 miles per 
second. These a rays thus move with velocities hundreds of 
times greater than the fastest moving meteor. Everyone 
is aware of the enormous energy possessed by a meteor 
moving, say, at 30 miles per second. But energy varies as 
the square of the velocity, and thus the a particle of radium 
possesses a quarter of a million times more energy, mass 


for mass, than a swiftly moving meteor. In this enormous 
energy of the rays lies the secret 6f the surprises of radium. 
From whence comes this enormous store of energy? 

In addition to its power of sending out radiations, radium 
possesses another important property, shared in by the radio- 
active substances actinium and thorium, namely, that of con- 
tinuously emitting a radioactive " emanation " or gas. This 
property is rendered very striking if a specimen of radium 
bromide is dissolved in water and the liquid evaporated 
down to dryness again to get the solid substance. This 
simple process has caused the radium to lose the greater 
part of its radiation. Strangely enough the radium slowly 
regains its activity, and if left entirely to itself, at the end 
of a month it is as radioactive as ever. Rutherford has 
showed that the solution in water causes the radium to give 
off a gas called " radium emanation." This emanation has 
all the properties of a true gas, it can be liquefied at a tem- 
perature of 150 C, but it is 100,000 times more radio- 
active weight for weight than radium. It does not combine 
with any known substance, and is not acted upon by any 
chemical reagent. It is not a radium compound, but it is a 
new element with an atomic weight which appears to be 222. 
It takes its place along with the rare gases of the air, argon, 
helium, neon, etc., and it gives a characteristic bright-line 
spectrum which shows neither the radium nor helium lines. 
It seems, therefore, that the element radium has been trans- 
formed into another element, radium emanation, or radon. 
If, after a month, the radium is again dissolved in water and 
evaporated to dryness as before, the radium loses its ac- 
tivity, and a fresh crop of emanation is produced. This 
same process may be repeated as often as possible with the 
result always the same, and we are perforce compelled to 
assume that the radium is continually manufacturing emana- 
tion, continually changing itself into a new element. This 
is really only the first of a series of changes, for radium 
emanation changes into radium A, and this in turn to radium 
B, and so on. This change is an atomic change going on 
within the atom. But how does this change progress? 

When the radium has given off the emanation, it still 


gives out a particles, but only about one-fourth as copiously 
as before the radium was put in water. The a particles are 
produced by the same change as makes the emanation, and 
the radium atom is therefore divided into emanation and 
a particle. 

Observations of the velocity and mass of the a particle 
made by Rutherford indicate either that the mass of the a 
particle is twice that of the hydrogen atom, or if the charge 
carried by the a particle is twice that of the hydrogen atom, 
then the mass of the a particle is four times that of the 
hydrogen atom and must therefore be an atom of helium. 
Hence each atom of radium apparently breaks up into one 
atom of helium and one of radon. 

All that was necessary to complete Rutherford's proof 
that helium was actually given off from radium, was to show 
experimentally that helium was thus produced. This was 
accomplished in 1903 by Sir William Ramsay and Frederick 
Soddy. A tube was filled with radium emanation which was 
separated from all other gases by condensing it with liquid 
air and removing by a pump the gas not condensed. This 
spectrum tube was sealed and the spectrum of the gas could 
be examined at will. At first no helium lines were visible, 
but after a lapse of three or four days, when the radium 
emanation had disintegrated, the spectrum of helium gradu- 
ally made its appearance, and finally the whole helium 
spectrum was complete. Similarly, Debierne has found by 
the spectroscope that helium is produced from the radio- 
active substance actinium, and Soddy has produced helium 
from uranium and thorium. Helium, therefore, has been 
found experimentally to be produced by the radioactive sub- 
stances radium, thorium, uranium and actinium. These sub- 
stances are alike in that each emits a particles. Hence, 
a particles are atoms of helium. Rutherford and Royds, 
however, have given a still more conclusive proof that the 
a particle is an atom of helium. These a particles are ca- 
pable of penetrating a certain small but definite thickness of 
glass. Glass may be blown very thin but yet retain its 
ability to remain air tight. Radium emanation was stored 
in such a thin-walled vessel and this enclosed in a second 


vessel. Alpha particles given off from the radium emana- 
tion thus could penetrate through the very thin glass walls, 
but were stopped in the outer vessel and were there col- 
lected. At first the gas in the outer vessel was found to 
contain no helium, but after some days, helium lines ap- 
peared in the spectrum, proving beyond a question of doubt 
that radium gives off helium. 

It is even possible to measure the rate of growth of the 
helium, which measures show that in a year, 168 cubic milli- 
meters of helium are spontaneously manufactured by each 
gram of radium. Rutherford and Geiger in this connection 
achieved one of the greatest triumphs for experimental 
science in being able to count the number of helium atoms 
or a particles that are ejected per second from one gram 
of radium. Indeed two different methods were devised 
which led to the same results. Both methods depend on the 
fact that each atom of helium as it is ejected gives a small 
flash like a meteor. By an electrical method, these flashes 
were counted by Rutherford and Geiger and it was found 
by them that thirty-four thousand million (3.4 x io 10 ) atoms 
of helium are ejected every second from each gram of 
radium. This number is in exact agreement with that ob- 
tained by noting with a microscope the number of scintilla- 
tions on a given area in a given time by the spinthariscope, 
invented by Sir William Crookes. Thus at the same time 
there was measured the amount of helium produced from 
radium, and likewise was given the number of molecules 
present in matter, information which was needed to com- 
plete many theories in physics. 

Investigations in radioactivity accordingly have given an 
entirely new conception of the atom. The atom is no longer 
one and indivisible, but certain atoms at least are trans- 
formed into other atoms, each radium atom being changed 
into one atom of helium and one of radium emanation. 
These atoms are continually changing, no less than thirty- 
four thousand million atoms of helium being produced each 
second of time from each gram of radium. As the atoms 
disintegrate, enormous stores of energy are let loose, and 
this energy manifests itself as light and heat. The heating 


effect of this energy has been measured and has been found 
to be 118 gram-calories per hour per gram of radium. A 
specimen of a grain of radium bromide would evolve about 
four calories per hour. In four years about 140,000 calories 
would have been evolved. An equal weight of coal would 
during complete combustion give out about 500 calories. 
Hence the radium in four years would give 280 times as 
much heat as if it had been coal and had been completely 
burned, and yet the radium in this time would diminish so 
very little in weight that it would be absolutely impossible 
to detect this diminution by the most sensitive balance 
known to modern science. The energy of radium comes 
from the disintegration of its atoms. The average life of a 
radium atom is 2280 years, so that in the complete life of 
one grain of radium about 100,000,000 calories are set free. 
This is 200,000 times more energy than if it were pure coal 
and entirely burned ! 

Helium, being permanent and not transitory, must ac- 
cumulate as the result of radioactive changes. In these 
changes, Soddy has shown the remarkable sensitiveness of 
the spectroscope in detecting slight quantities of helium 
for he has proved, in numerous special experiments, that the 
D.< line of the helium spectrum can be detected with cer- 
tainty, if only one millionth of a cubic centimeter, or one 
five-thousand-millionth part of a gram of helium is present. 

An achievement of far-reaching importance in all theories 
of physics was the discovery by Sir J. J. Thomson of a 
body having a mass much less than that of the lightest 
known atom, hydrogen. This body, called by its discoverer 
a corpuscle, but now known as an electron, is i/i845th part 
of the mass of the hydrogen atom. A further study of the 
electron showed that it is always associated with a negative 
charge of electricity and in fact carries a unit charge of 
negative electricity. The physicist has now become con- 
vinced that the atom is an electrical structure made up of 
nearly equal amounts of positive and negative electrical 
charges. The atom is believed to consist of a central group 
of elementary positive charges, or protons, with a smaller 
number of negative charges, or electrons, called the nucleus, 


and about this nucleus there is an outer system of negative 
electrons, varying in number from one to ninety-two. These 
outer electrons can be expelled from the atom by a number 
of different methods, such as the application of heat, impact 
of ions, exposure to ultra-violet or X-rays, or they may be 
emitted by radioactive substances; in fact, the (3 rays con- 
sist of a stream of negatively charged particles. 

The discovery by C. T. R. Wilson that the charged ions 
produced in gases by a and p rays become the centers for 
the condensation of water vapor paved the way to experi- 
mental work of a remarkable nature. Millikan has secured 
extraordinary results by utilizing tiny drops of oil in place 
of water, and as a result of his experiments, he has been 
able to prove conclusively that the electrical charges carried 
on ions " all have the same value or else small exact mul- 
tiples of that value. " This fundamental unit is the same, 
both for positive and negative electricity, and is numerically 
equal to the charge carried by the negative electron. This 
unit charge of electricity was measured by him to an ac- 
curacy of one part in 1000. With this information, it was 
possible to estimate with greater accuracy the mass of the 
electron in grams and the number of molecules of any gas 
per cubic centimeter at o C and 760 mm pressure. 

Since we know therefore the size of a molecule and the 
number of molecules per cubic centimeter, it is possible to 
compute the number of molecules through which the a or 
the (J rays emitted by radium must pass in going a given dis- 
tance. The extraordinary fact revealed (by the photographs 
of Wilson, referred to above) is that the swift-moving (J par- 
ticles pass, on the average, through as many as 10,000 atoms 
before coming close enough to an electron to detach it from 
its system and form an ion. In fact, it has been shown by 
Eddington L that when an electron encounters an ionized 
atom it will be captured if, and only if, it actually hits the 
nucleus of the atom. The electron must therefore form 
but a very minute portion of the space enclosed within the 
atomic system. The a particle being an atom of helium with 
a mass more than seven thousand times that of the negative 

1 Monthly Notices, R. A. S., 83, 32, 1922. 


electron, it cannot be deflected from its course by an elec- 
;ron which is of very minute mass, but only by some ponder- 
ible mass at least comparable with that of the helium atom. 
This heavier mass is found at the nucleus of the atom. As 
i result of Geiger and Marsden's experiments on the scat- 
ering of a rays, it was found that when these rays passed 
hrough very thin metallic foils, the deflections witnessed 
:ould be explained only by assuming a very close approach to 
i small but massive charged particle. Rutherford x was 
tccordingly led to assume that the typical model of an atom 
onsisted of an exceedingly minute and comparatively mas- 
ive positively charged nucleus, about which is collected a 
lumber of electrons. Each and every one of the electrons 
orming the outermost parts of all atoms are exactly alike 
,nd each carries a unit amount of negative electricity. Since 
ach atom is electrically neutral, the charge on the positive 
lucleus must be equivalent to the sum of those carried on the 
J electrons. The value of N for each of the atoms is a f unda- 
lental constant, for on it depends the size of the electric 
ield surrounding the nucleus and the peculiar arrangement 
f the external electrons, which in turn determine the physi- 
al and chemical properties of the atom. Experiments in 
911, by Barkla on the scattering of X-rays indicated that 
he number, N, of electrons in an atom was approximately 
alf the atomic weight of the element. This conclusion was 
bundantly verified by the magnificent work of Moseley. 
le found that the X-ray spectrum was similar for all ele- 
lents, and that when he plotted the square root of the 
requencies of the characteristic X-ray spectra, all the ele- 
lents examined arranged themselves upon nearly perfectly 
:raight lines. The atoms were then numbered in the order 
i which these spectra placed them to give these straight 

The ordinal number corresponding to the place occupied 
y each element in the periodic table (p. 306) has been 
>rmed its atomic number. The work of Moseley showed 
lat all the chemical elements took their proper places in 
le periodic table, and as uranium is the heaviest atom 

1 Phil. Mag., 21, 669, 1911, and 27, 488, 1914. 


known (atomic number 92) there can be only 92 species of 
elements. With elements 85 and 87 now thought to be 
known, all gaps have been filled in the periodic table. 

These discoveries of the physicist have been the greatest 
boon to the work of the chemist. The latter has always had 
great faith in the principles underlying the Mendeleeff table, 
but in this table, arranged in order of increasing atomic 
weights, certain discrepancies appeared; for instance, argon, 
of atomic weight 39.88, from its properties was compelled 
to find a place in the table before potassium having a smaller 
atomic weight of 39.1. The system of atomic numbers 
places A with number 18, in its rightful place before K with 
atomic number 19. In a similar way, the system of atomic 
numbers places Co in the table before Ni, instead of after 
it, if arranged according to atomic weights. 

With a knowledge of atomic numbers, the difficulties of 
classification presented by radioactive substances were now 
cleared away. Soddy has shown that when an a particle is 
emitted, the position of the element in the periodic table 
is shifted by two numbers to the left towards smaller num- 
bers, while if a [4 particle is emitted the atomic number in- 
creases by one unit, and the place is shifted one to the right. 
Since the a particle is an atom of helium with positive charge 
2 and mass 4, while the (3 particle is a negative electron with 
no appreciable mass, it is evident that the emission of an a 
particle will diminish the atomic weight by 4, but the emis- 
sion of a (J particle will cause no change in the atomic weight. 
Consequently, if the emission of an a particle by a substance 
is followed by two successive changes in which (J rays are 
set free, the net result will be that the element, after ex- 
periencing these three changes, will move back again into 
the position in the periodic table it had held originally. 
These changes have not altered the size of the atomic num- 
ber, but have diminished the atomic weight by 4, and con- 
sequently it is possible for two or more elements to have 
the same atomic number but differ in their atomic weights. 
Such elements are known as isotopes. Isotopes are indistin- 
guishable from each other by any chemical tests, or by any 
spectroscopic tests since the spectra are identical. Radium 


(at. wt 226), Th X (at. wt. 224) and Ac X (at. wt. 222) 
are examples of isotopes, each possessing a nuclear charge or 
an atomic number of 88. 

Since all masses are nothing more than electromagnetic 
manifestations, and since the mass of the electron is very 
minute and negligible compared with the mass of the nu- 
cleus, it should be possible to compare the masses of dif- 
ferent elements by subjecting them to successive electric and 
magnetic fields. This method of " positive ray analysis " 
developed by J. J. Thomson consists in measuring the ratio 
of charge to mass. This was made possible by a beautiful 
photographic method, which in the capable hands of Aston 
has greatly improved our knowledge regarding the atoms. Al- 
though demanding technical skill of a very high order, the 
chemical examinations can be carried on by simple methods 
leading to definite results. A small supply only of the gas 
to be investigated is required ; it need not be chemically pure 
nor need any special care be taken to wash away from the 
vacuum tube all traces of the last gas investigated. Ac- 
cording to Aston, it is impossible to remove from a tube " all 
visible traces of a misspent career." Until quite recently 
it was thought that the experimental measurements proved 
that the elements whose atomic numbers are whole num- 
bers, with oxygen assumed at 16.00, are pure elements, while 
all other elements with fractional atomic weights are mix- 
tures of isotopes, each of the isotopes however having a 
whole number for its atomic weight. B (at. wt. 10.9) is a 
mixture of two isotopes of masses n and 10; Ne (at. wt. 
20.2) consists of two isotopes, masses 20 and 22; Mg (at. 
wt. 24.32) is a mixture of three, of masses 24, 25 and 26; 
Cl (35.46) consists of two, of masses 35 and 37; while Xe 
and Hg are each made up of no less than 6 isotopes. This 
conclusion has been greatly modified by the discovery that 
each of the elements C, H and O has isotopes. 

Our view is that the nuclei of all atoms are made up of 
multiples of hydrogen nuclei each carrying unit positive 
charge, the combination being bound together by the exter- 
nal electrons. Hence the mass of the atom is confined to the 
nucleus, but the size of the nucleus is very minute compared 


with the whole volume of the atom. In fact, the radius 
of an electron cannot be larger in comparison with the radius 
of the atom than is the radius of the earth compared with 
the distance from earth to sun. Each atom therefore forms 
a miniature solar system, the external electrons being held 
in place and compelled to perform their orbital motions by 
the comparatively massive nucleus. Since there may be as 
many as 92 external electrons, it is evident that modern 
mathematics cannot furnish a general solution of the motions 
of the electrons, except in the case of the very simplest of the 

Since the chemical and physical properties depend on the 
distribution of the electrons of the outer atom, there have 
been many attempts to formulate a structure for the atom. 
From such attempts have gradually evolved atomic models 
of many different types. The Lewis-Langmuir atom has 
been very successful in explaining the chemical properties, 
particularly the valence. This atom is not based on any 
dynamic principles. Valence, which measures the power to 
combine, may be positive or negative depending on whether 
the atomic system has too many or too few electrons to 
make a stable combination. The inert gases helium, neon, 
argon, krypton, xenon and radon are elements which have 
no power to form compounds. Such atomic systems under 
ordinary conditions cannot capture an electron from an- 
other atom, nor can they get rid of one of their own. The 
inert gases have atomic numbers of 2, 10, 18, 36, 54 and 86, 
and hence we may imagine the atoms as if made up of con- 
centric shells of electrons, the shells containing 2, 8, 8, 18, 
1 8 and 32 electrons respectively. 

Except for the hydrogen system, all atomic structures 
have the two electrons forming the system of helium as 
their innermost shell. Since helium is inert and cannot take 
up another electron; the third electron which forms lithium 
must be a single electron in a shell exterior to the helium 
system. The lithium system is therefore not very stable, 
and readily gives up its external electron. It has a positive 
valence of one. The system of fluorine, with seven electrons 
in the second shell, may be regarded as being in a position 


readily to capture an electron from an atom in its neigh- 
borhood, and so it has a positive valence of seven or negative 
valence of one. Likewise oxygen may be regarded as having 
positive or negative valences of 6 and 2 respectively. The 
Lewis-Langmuir atom is very successful in explaining the 
relative positions of the elements in the periodic table. 
The electrons forming the atom of this model are relatively 
fixed in position, for though each electron may be in motion, 
it is confined to a small portion of the space occupied by the 
atom. With the advance in knowledge this atomic model 
has been superseded by others. 

Great difficulties, however, appear when discussing atomic 
models on the principles of physics, especially when there is 
a transference of energy from place to place, or when the 
motions of the atoms make them depart from the steady 
state thereby causing emission or absorption of light. After 
repeated failures to explain these matters by means of the 
accepted theories, Planck made one of the most startling 
proposals ever presented to the scientific world, from which 
developed his celebrated quantum theory. According to 
Planck, it was assumed that the transference of energy can 
only take place in definite but very small units, and that 
the total energy transferred is always an integral multiple 
of this small unit, called the energy quantum. This can 
be expressed simply in mathematical terms. If Ei and E 2 
represent the energy of a system before and after radiation 
has taken place, then the energy spent in radiation is 
Ei E> = hv, where h is Planck's constant and v is the fre- 
quency of vibration of the body concerned. This mathe- 
matical equation gives expression to the simple statement 
that the total amount of energy emitted or absorbed by a 
radiating body is always proportional to the frequency of 
vibration, which in turn is inversely proportional to the 
wave-length of the light emitted or absorbed. 

The very radical nature of Planck's hypothesis may be 
estimated when it is stated that it stands at variance with 
all the previously known laws of mechanics developed to 
explain the motions of material objects of large dimensions. 
As a matter of fact, there is no absolute necessity that the 


same laws that apply to ponderable material in a gross state 
must also be applicable to simple atoms. This duality of 
laws, a topic of much heated discussion by physicists in re- 
cent years, has given rise to the expression " classical dy- 
namics " which explains the motions of matter in obedience 
to the law of gravitation. One fundamental conclusion of 
the quantum theory is that motion is not the continuous 
process that we have accustomed ourselves all our lives to 
believe; but the motion takes place " steadily by jerks/' 
the jerks however being so small that the process is to all 
practical purposes continuous. 

Of the many atomic models proposed, the most successful 
in explaining the physical facts, and more particularly the 
spectroscopic data, is the Bohr-Sommerfeld atom. This was 
first proposed in 1913 and it is based on Planck's quantum 
theory of energy. In explaining the motions of the external 
electrons, evidently little hope can be expected from the 
classical dynamics, for this would require the solution of the 
problem of n bodies when n is large. For the present, as- 
sumptions must be made for the purpose of securing results, 
and in spite of much inconsistency, the quantum theory of 
spectra is the most satisfactory attempt so far made to- 
ward interpreting spectral series. In the simple case of the 
hydrogen atom and ionized helium, each with one external 
electron, the Bohr-Sommerfeld method has been very suc- 
cessful in reproducing many of the details of the spectra 
both in electric and magnetic fields. On the basis of this 
theory, the problem of atomic structure consists in building 
up each atom in such a way that the passage of an electron 
from one stationary state to another will give the observed 
wave-lengths. Although the acceptance of the principles of 
wave mechanics gives a better explanation of the theory of 
spectra, nevertheless the Bohr atomic model is still useful 
in explaining the facts. 

In the investigations of spectra, the first of the lines to 
show an arrangement in series were those due to hydrogen, 
discovered in 1880 by Huggins in the spectra of white stars. 
This series is an extension of the four lines visible in the 
solar spectrum. A new era in spectroscopy started in 1885 


when the law underlying the hydrogen series was discovered 
by Balmer. The thirty-five lines found in the flash spectrum 
are represented by the formula 

X =3646.125-^- 

where X is the wave-length on Rowland's scale and m takes 
the values 3, 4, 5. ... 

Shortly afterwards, Kayser and Runge, and Rydberg in- 
dependently, began the publication of their splendid re- 
searches. Rydberg's investigations are of the greatest im- 
portance since they have laid the foundation for all future 
work on spectra in series. He began by sorting out doub- 
lets and triplets and thus ascertaining the lines which belong 
together in a series. He was able to distinguish three chief 
kinds of series, as follows: 

Principal, including the strongest lines 
Diffuse, of intermediate intensity 
Sharp, including the weakest lines. 

Each of these three series may consist of single, double or 
triple lines. Each and every series always converges to- 
wards a limit at short wave-lengths, and the lines at the 
same time diminish in intensity. A fourth so-called " fun- 
damental " series with lines mainly in the infra-red has been 
discovered by Bergmann. 

Many and varied are the mathematical formulae em- 
ployed to represent the series of spectra. The most satis- 
factory formula, which is due to Rydberg, takes the following 

A N 

V m = A - 

(m + ju) 2 

where A is the limit of the series, N is the " Rydberg con- 
stant" for hydrogen, and the wave numbers v m are ob- 
tained by assigning successive integral values to m. \i may 
be regarded as a decimal part to m, though it is sometimes 
greater than unity. For the details of the investigations of 


series spectra, one would do well to read Fowler's excellent 
Report on Series in Line Spectra, 1922. 

Of special interest in dealing with the flash spectrum are 
the investigations regarding enhanced lines. Fowler has 
shown l that these lines form series entirely similar to those 
of the ordinary lines. The formula representing the en- 
hanced series, however, differs from that of the ordinary 
series in that the Rydberg constant N is multiplied by 4. 
This has a simple explanation from Bohr's theory. The or- 
dinary series lines are emitted when an electron of charge e 
returns to an atom from which it has been displaced. The 
enhanced lines, on the other hand, are produced when an 
electron returns to the atom which has already lost another 
electron through ionization; consequently, two electrons 
are detached from the atom, each electron carrying a charge 
e, or a total charge 2e. The formula for the Rydberg con- 
stant TV, involves the square of the charge e, and hence for 
the enhanced lines the multiple 4 appears. Enhanced lines 
therefore belong to the ionized atom, or one which has lost 
a negative electron and hence carries an excess of positive 
charge. According to a suggestion by Saunders, He-\- and 
C0+, refer to ionized helium and calcium, the addition of 
the + sign following the chemical symbol signifying that the 
atom is not electrically neutral but carries a unit + charge. 
After an atom has lost one electron and thus becomes ion- 
ized, it may lose a second electron and become " doubly 
ionized." The symbol adopted for calcium under these con- 
ditions is Ca++ ? the atom carrying two extra positive 
charges. According to Bohr's theory, the charge concerned 
in the production of spectrum lines by such an atom is 30, 
and hence the Rydberg constant TV must be multiplied by 9. 

This is not the place to give the mathematical theories 
underlying the formation of spectrum lines, but a synopsis 
may be given of the more important developments found in 
Fowler's Report on Series in Line Spectra, 1922, and in 
Russell, Dugan and Stewart's Astronomy, 1927. Any ac- 
cepted theory must explain why the frequency of any line in a 
spectrum appears always as the difference between the terms 

1 Phil. Trans. A. 214, 225, 1914. 


of a quantity, neither of which represents a spectral line, 
and must furnish an explanation of the physical meaning 
of the two terms, and must further explain how an emitted 
frequency comes to be the difference of two of these. 

Adopting the idea of Rutherford, outlined above, that 
each atom consisted of a heavy nucleus carrying a positive 
charge which was surrounded by negative electrons, Bohr 
was able to give a satisfactory theory in the case of the sim- 
plest type of atom. The hydrogen atom is such a unit since it 
consists of a single electron in orbital motion around the nu- 
cleus, and equally simple are the atoms of enhanced helium 
and doubly enhanced lithium. According to the Bohr 
theory, the single external electron is free to traverse cer- 
tain specified orbits, which are determined in the simple 
case of circular orbits by the condition that the angular 
momentum is an integral multiple of h/2 IT, where h is 
Planck's constant, derived from the quantum theory. When 
the motion of the electron is confined to one of these sta- 
tionary orbits, there is no radiation. Emission occurs 
only when the electron passes from one stationary orbit to 
another. Without attempting to explain the mechanism 
which causes the electron to pass from orbit to orbit, Bohr 
supposes that the transition is followed by the emission of 
light, the frequency of which can be determined from the 
quantum theory. In fact, the energy radiated is equal to 
the differences of the energies of the electron in the two 
orbits concerned, and is assumed to be one quantum of 
energy, hv, where h is again the Planck constant and v 
is the frequency. At a given instant, any one electron falling 
from an external to an internal orbit causes one line only 
in the spectrum, and it is the summation of the actions of a 
large number of electrons that causes the whole series of 
spectrum lines. 

By the elementary laws of mechanics, it is possible to de- 
rive the necessary equations for orbital motion. Taking 
E and M as the charge and mass of the nucleus, e and m as 
the charge and mass of the electron, c the velocity of light, 
and v the wave-number of the line, then for the case of 



2?r z Jf e (i i 

v = m(-~- - 

ch* W / 2 2 

where ti and /2 are integers. 

This formula is of exactly the same form as that which 
represents the Balmer series of hydrogen, the quantity out- 
side the brackets representing the Rydberg constant. In 
fact, as a splendid confirmation of Bohr's theory, the Ryd- 
berg constant, N, calculated from Millikan's data, furnishes 
a value which agrees with that found from spectral series 
within an accuracy of one part in 1000. 

FIG. 4 Orbits of the hydrogen atom. 

The successive orbits of hydrogen are in the ratios r, 2 2 , 
3 2 . . . In the normal state of the atom the single electron 
revolves in the innermost orbit. When the atom is disturbed 


ionized helium have been measured by Plaskett in certain 
O-type stars. At the position of HOL the difference in wave- 
length from the hydrogen to the helium line is 2.63 A, while 
at H6 this diminishes to 1.54 A. 

It has thus been found that the enhanced spectrum of 
helium resembles that of neutral hydrogen, and in exactly 
similar manner it has been concluded that in all details (even 
in showing doublets or triplets in their spectra) the enhanced 
spectrum of an alkali earth (like Mg) resembles the arc 
spectrum of the alkali metal of next lower atomic number 
(like No). It has further been concluded that this relation 
exists for all alkali earths and alkali metals. In fact, this 
similarity in spectra between neighboring elements in the 
periodic table appears to be a general rule, namely, that the 
enhanced spectrum of any element resembles the arc spectrum 
of the element of next lower atomic number. If an element 
loses two electrons, and is therefore doubly ionized, its spec- 
trum, for the same reason, should be similar to the arc spec- 
trum of the element of second lower atomic number. Fowler 
has succeeded experimentally in discovering doubly- and 
also trebly-ionized silicon. Interesting comparisons have 
been made by him between the spectra of trebly-ionized Si, 
doubly-ionized Al, singly-ionized Mg and neutral Na, i.e., 
between the spectra of four succeeding elements in the pe- 
riodic table of page 306. In the laboratory Millikan almost 
at will has been able to strip external electrons from atoms. 

Although the mathematical analysis for a generalized 
theory has been too difficult to follow through to completion, 
nevertheless the Bohr theory has been successful in placing 
spectrum analysis on a very firm foundation. Further de- 
velopments will be treated in the following chapters. 



HALF a century of remarkable progress in solar 
physics since Young first observed the flash spectrum 
at the eclipse of 1870 saw a vast diversity of opinion 
regarding conditions of temperature and pressure in the 
envelope causing the dark Fraunhofer lines. On the one 
hand appeared Young's original view that the reversing 
layer is a thin shell but a few hundred miles in thickness, 
while on the other extreme we had Lockyer's opinion that 
such a reversing layer does not exist and that the corona is 
merely the cooler and rarer portion of the chromosphere. 
Within the past decade, however, the enormous increase in 
knowledge concerning the structure of the atom, explained 
in the preceding chapter, coupled with the publications of 
discussions of the spectrum of the chromosphere obtained at 
solar eclipses, has resulted in the solution of many of the 
outstanding problems relating to the sun. 

In spite of the enormous improvement in recent years in 
instrumental equipment and in technique, it is still no exag- 
geration to state 1 that it is more difficult to secure a per- 
fectly successful photograph of the spectrum of the chromo- 
sphere than it is to obtain an excellent photograph of any 
other single phenomenon attacked by astrophysical science. 
Witness the fact that since its first observation in 1870, the 
flash spectrum has continually been and still remains one of 
the most important of all problems taken up for solution at 
each succeeding solar eclipse. It is now nearly forty years 
since the first photograph of the flash spectrum was ob- 
tained, but in this interval of time, comparatively long when 
judged by the attainments of modern science, there have 

1 Cf. Handbuch der Astrophysik, 4, 275, 1929. 


been more than one hundred attempts to photograph the 
flash spectrum; and yet there are not more than a half 
dozen photographs which may be considered to rank as first 
quality. Of course, it is perfectly true that many of the 
attempts might have succeeded had it not been for clouds, 
or thin haze, or poor atmospheric conditions, yet the fact 
remains that the great majority of those photographs which 
escaped the clouds have suffered from lack of perfect defini- 
tion or from inaccurate timing of the exposures. 

At an eclipse, the flash spectrum may be observed visually 
or photographically, either with or without slit, and by means 
of prisms or gratings. The researches of the present day 
demand large dispersion. The type of spectrograph depends 
very largely on the particular problem which the eclipse ob- 
server desires to solve. The great advantage of the prism 
is the greater light-gathering power, the light being concen- 
trated in one spectrum, but on the other hand the grating 
possesses many points in its favor. The lines in the prismatic 
spectrum are crowded together at the red end and widened 
out at the blue part of the spectrum, thus entailing much 
difficulty in the determination of wave-lengths. The grating 
gives a normal spectrum, permits of higher dispersion, and 
gives higher resolving power, a larger extent of spectrum 
and probably better definition. Gratings either plane or 
concave may be used, but with a flat grating a lens becomes 
necessary to bring the spectrum to a focus, and such a lens 
introduces aberrations and absorption of light, and conse- 
quent loss of definition. In the past, plane gratings have 
been used at eclipses, but for the future it is safe to predict 
that whenever gratings are employed for the chromosphere 
they will be concave gratings and not plane. 

The spectrograph which lends itself most readily to the 
easiest and sharpest focus under the temporary conditions 
of eclipse observations is undoubtedly one used with slit. 
For this purpose it is possible to take bodily into the field a 
spectroscope of the type that has been thoroughly tested 
and tried out on stars in the regular work of the observatory. 
This instrument is universally a prism spectograph with 
prisms of the very highest quality. It is readily possible to 



W ctf 

P. u 


a u 

O J3 

K tJ 

a 'S 

u * 






a. = 


2 - 

w W 






test thoroughly the spectograph at home and then to mark 
carefully the positions of the focus for collimator and camera 
before dismounting. It is the work of a short while to again 
assemble the spectrograph at the eclipse site, and since the 
whole spectrograph is comparatively light in weight it can 
easily be picked up and set down anywhere. For instance, 
it is readily possible to test the focus thoroughly by the use 
of the electric arc, or by taking a focus plate of the daylight 
sky. At the time of the eclipse such an instrument is ordi- 
narily used with a heliostat or coelostat mirror, and with the 
employment of a lens or concave mirror for forming an image 
of the sun on the slit of the spectrograph. With this type 
of spectrograph an eclipse observer should be absolutely 
certain to secure spectra of the chromosphere in excellent 
definition at each and every eclipse where clouds do not 
interfere. The only chances of failure with such an instru- 
ment are poor focus of the spectrograph and inability to 
keep the portion of the image of the sun, selected for observa- 
tion, on the slit at the time of the eclipse. Any observer 
skilled in observatory manipulation through long years of 
handling a slit spectrograph should be ashamed to admit 
that the focus was poor due to the lack of careful methods 
of adjustment. There is however some excuse if a skilled 
astronomer fails to keep the slit of the spectrograph filled 
with light for the reason that the few moments of an eclipse 
are very tense and excited, the light at totality is feeble, and 
after all, even an efficient observer must at some time in his 
career see his first total eclipse and he will have few oppor- 
tunities of witnessing others. 

By the use of slit spectrographs, the British observers have 
obtained excellent photographs of the flash spectrum. At 
the 1926 eclipse observed in Sumatra, Davidson and Strat- 
ton * used a quartz spectrograph designed to photograph to 
the extreme limit of the ultra-violet. The prism train con- 
sisted of four double quartz prisms of 60 angle, each prism 
being composed of two half-prisms of right- and left-handed 
quartz. The collimator and camera objectives were each 
a single quartz lens of 3 -inches aperture and 36-inches focus. 

1 Memoir R. A. S., 64, 105, 1927. 


Excellent definition was secured between wave-lengths 3066 
A and 4200 A. 

The greatest dispersion ever used at an eclipse, before 
that of 1932, was employed by Pannekoek and Minnaert 1 
in 1927. Their optical system consisted of many elements 
each of unusually large dimensions. After reflection from 
the coelostat mirror, the rays passed first through an image 
lens 10 inches in diameter. After passing through a slit, a 
totally reflecting prism turned the beam through 90, while 
next in the train came a triple collimating lens of 6-inch 
diameter. After traversing three prisms of 45 angles and 
large enough to receive the full beam of light, the rays were 
reflected from a plane mirror, and then they passed in re- 
versed order through the prisms and 6-inch objective, which 
now became the camera lens for forming an image on the 
photographic plate. Good definition was obtained in the 
region of wave-lengths 4153 to 4751 A. 

Although it is easier to photograph the flash spectrum by 
using a slit, nevertheless a slit is entirely unnecessary, the 
crescent arcs of the chromosphere providing narrow curved 
lines in the photographed spectrum. Good reasons for dis- 
carding the slit will be given later. However, it is a much 
more difficult task to secure perfect focus of the chromo- 
spheric spectrum without a slit, either by the use of objective 
prisms or grating. Slitless spectrographs of the power essen- 
tial for the flash spectrum are practically never used in 
observatory or laboratory. In stellar work, a prism is placed 
in front of the objective, forming what is universally called 
a " prismatic camera. " For many long years this type of 
instrument has been used in routine observatory work, but 
ordinarily the dispersion is small and therefore unsuited to 
making any contribution to our present knowledge of the 
flash spectrum. The dispersion of such a combination may 
be increased by two different methods: (i) by increasing 
the focal length of the objective; or (2) by increasing the 
number of prisms. At the 1926 eclipse, Davidson and Strat- 
ton used an ingenious device, the camera of ig-foot focus 
being utilized in a dual capacity. With a direct- vision slipped 

1 Verh. Kon. Akad. ie Amsterdam, 13, No. 5, 1928. 


in front of the objective the flash spectrum was photographed, 
while with the prism removed, direct photographs of the 
corona were obtained. Even with the moderate focal length 
of 19 feet, necessary to secure the desired dispersion, the 
crescent arcs were two inches in diameter on the photo- 
graphs. As will be explained later, the average conditions 
of seeing experienced at eclipses may cause these crescents 
forming the lines of the spectrum to lose their clear-cut char- 
acter in spite of the fact that the focus of the instrument 
may be perfect. None of the half dozen photographs of the 
flash spectrum which the writer regards as of the highest 
class of excellence have been obtained by this type of in- 
strument, where the increased focal length of the camera has 
been relied upon to increase the dispersion. 

On the other hand, excellent spectra have been obtained 
by objective prisms without the use of slit, through the em- 
ployment of additional prisms to increase the dispersion. 
At the eclipse of 1905 in Spain, Campbell 1 secured exquisite 
definition with two 60 prisms in front of an objective of 
2 -inch aperture and 6o-inch focus. Lines were recorded 
between wave-lengths 3820 and 5300 A, though the defini- 
tion began to fail in the neighborhood of 4650 A. The dis- 
persion at Hy was 5.7 A per mm. A moving plate was used 
in forming the spectra, a method which will be explained 
later. At the Flint Island eclipse of 1908, Campbell in- 
creased the dispersion by the use of three 60 prisms and 
an objective of 60 inches focal length, the dispersion being 
about double that obtained by him in 1905. At Hy this 
amounted to 2.9 A per mm, almost as great in scale as in the 
1927 spectra of Pannekoek and Minnaert. 

In the observation of eclipses, my own work has always 
been spectroscopic. I have always employed gratings and 
have used them without a slit. My record as an eclipse ob- 
server is probably unique. I have observed eight solar eclip- 
ses, and yet neither the Leander McCormick Observatory 
nor I, personally, owns a single piece of eclipse apparatus; 
at each eclipse everything must be borrowed. At my latest 
eclipse, in 1930 at Niuafoou Island, I had two concave grat- 

1 Publications of the Lick Observatory, 17, i, 1931. 


ings. All of the equipment used at this expedition belonged 
to the U. S. Naval Observatory with the exception of one 
grating of 4-inch aperture kindly loaned by my good friend 
Saunders of Harvard University. This grating has been 
with me at every eclipse since 1905. Photographs of the 
flash spectrum were obtained in all but two of the eclipses, 
in 1923 in California and in 1927 in Norway. The results 
from the photographs obtained under clear skies in 1905 
and 1925 have been published in Astrophysical Journal, 71 , 
i, 1930, and 72, 146, 1930. The other concave grating used 
in 1930 was of 6-inch aperture. Both gratings are of lo-foot 
radius of curvature (or 6o-inch focus) and each has approxi- 
mately 15,000 lines per inch. The dispersion is 10.9 A 
per mm. At this recent eclipse, excellent definition was 
obtained from 3200 in the violet to 7800 A in the red, 
the region 4650 to 6800 A being photographed by both 
spectrographs. The discussion of these spectra has not been 
completed. The same two gratings will be employed at the 
1932 eclipse. 

The arrangement for using the concave grating without a 
slit is one of the greatest simplicity. The light from the sun 
falls directly on the coelostat mirror and then after reflec- 
tion the beam of parallel light falls on the grating where it is 
diffracted and brought to a focus on the photographic plate. 
If the grating and the photographic plate are each perpen- 
dicular to the line joining their centers, the spectrum is nor- 
mal, or to speak in more exact terms, the spectrum departs 
very little from a uniform scale of wave-lengths. 

Used in the ordinary Rowland form of mounting in the 
laboratory, one of the well recognized advantages of the 
concave grating is the property of " astigmatism/' whereby 
the spectrum lines are increased in length. If the astigma- 
tism should be of approximately the same amount when the 
grating is used objectively without slit, then as a result of 
lengthening out the chromospheric lines, which are neces- 
sarily curved, the definition would be ruined. Consequently, 
in making plans for 1900, when concave gratings were used 
for the first time at an eclipse, the Naval Observatory party 
did not dare attempt to use such a grating without slit. 


The successful photographs of stellar spectra l secured by 
concave grating used objectively showed however that these 
fears were groundless. Moreover, Runge's discussion of the 
theory of the concave grating, in Kayser's Handbuch der 
Spectroscopie, i, 450, 1900, proved that the amount of as- 
tigmatism for a concave grating used in the objective form 
would be so minute that it could have no harmful effects 
on the definition of the spectra. 

It should hardly be necessary to add that the grating 
spectroscope must be very firmly mounted on solid piers of 
masonry or heavy timbers in order that the tremors of the 
apparatus caused by the wind or by the changing of the 
plate holders may quickly subside. It is manifestly difficult 
to mount such an instrument of large dispersion on an equa- 
torial mounting or on a polar axis, with the grating in conse- 
quence directly exposed to the sun's rays. This method 
would indeed get rid of the coelostat mirror with its possible 
change in figure, but if this plan were followed, it would 
probably be a case of " out of the frying pan into the fire." 

It might not be out of place to call attention to the very 
great difficulty always experienced by eclipse observers in 
securing sharp focus with their spectroscopes; as stated, one 
very prominent feature of eclipse spectra in the past has 
been the continued succession of photographs poorly fo- 
cused. One method of securing focus frequently made use 
of has been to apply the final adjustments by utilizing the 
spectrum of the disappearing crescent of the sun a few 
minutes before totality. During the excitement and nerve- 
racking tension of these moments, a perfect adjustment can 
be obtained in this manner only as the result of a happy acci- 
dent and this method should never be resorted to under 
any circumstances. For instruments of small dispersion, the 
light from a star may be utilized for securing focus, if hap- 
pily some bright star is conveniently located, but for spectro- 
graphs of the greatest dispersion the spectra even from the 
brightest of the stars are too weak. For large instruments, 
there is left only the option of securing focus on the sun itself 
or the electric arc several days before the eclipse by the em- 

1 Astrophysical Journal, 10, 29, 1900. 


ployment of some sort of collimating device which will give 
a parallel beam of light coming from a slit source. For the 
use of the U. S. Naval Observatory party at the eclipse of 
1905, Jewell constructed a collimator consisting of a slit at 
the common focus of two concave mirrors, lenses not being 
used because of their chromatic aberrations. Several meth- 
ods of placing the slit at the common focus of the mirrors 
will at once suggest themselves to any ingenious eclipse ob- 
server, one of the simplest being to utilize a telescope of me- 
dium size (say of five inches aperture) accurately focused on 
the stars. 

The author has always made it a habit to be at the eclipse 
site a month or more in advance so that there may be plenty 
of time to secure exact focus without too much rush and 
excitement immediately preceding the eclipse. The collima- 
tor described above has always been utilized. When electric 
power is available, it has been found easier and more efficient 
to use the electric arc as the source of light for adjusting 
rather than the sun. This method was followed in the 
eclipses of 1923, 1925 and 1927. At the " Tin-Can Island " 
eclipse there was no electric power, and hence there was noth- 
ing to do but use the tropical sun as the source of light during 
the weeks of adjustment. The difficulties underlying this 
process are at once evident. In the optical train while ad- 
justing, there are four mirror surfaces: the coelostat of 
silver-on-glass, the two metal mirrors of the collimator and 
the grating itself. In the adjustment, the slit becomes the 
source of light, and hence it is of minor importance if the 
coelostat mirror is not plane or if the slit is not in the exact 
focus of the first collimator mirror which forms the image 
of the sun on the slit. 

Unfortunately, mirrors alter their focal lengths with 
changes in temperature, a fact well known to every astrono- 
mer with experience with a reflecting telescope. Excellent 
focus of the flash spectrum requires, among other things, that 
on the eclipse day the coelostat mirror be plane and not 
warped by the direct rays of the sun. (The mirror is always 
screened from sunlight until a few minutes before totality.) 
No matter how much time has been spent nor how much care 


in the determination of focus, the great fear is always present 
lest the grating alter its focal length owing to changes of 
temperature between the time of adjustment and the time 
of the eclipse. This requires that the wooden boxes and 
plate holders of the spectrograph should not expand enough 
to deteriorate the focus. And all of these nice adjustments 
had to be made and relied upon at the recent eclipse on a 
tropical island, with clouds always interfering with the work, 
with the dark-room during the daytime with temperatures 
that felt like inferno, and with no ice and no running water 
for development of the photographs. The trials and tribu- 
lations of an eclipse astronomer are many, especially when 
it is necessary to work with apparatus which is temporary 
and always borrowed ! 

With the collimator placed in the beam of light between 
coelostat mirror and grating, the focus was first determined 
visually with as great care as possible. For improving the 
focus, about a hundred and fifty focus plates were taken. 
The extreme range of focus in the last hundred of these films 
was but a fiftieth of an inch together with slight changes in 
the tilt of the photographic film. 

The best type of spectograph to plan for the photography 
of the flash spectrum will unquestionably be the one that is 
able to make the greatest contribution to the solution of solar 
problems. With the gigantic equipment of great observa- 
tories working assiduously on the dark-line absorption spec- 
trum of the sun, with the knowledge of the structure of the 
atom gleaned in recent years from refined researches in ob- 
servatory and laboratory, how can the bright-line emission 
soectrum of the sun photographed with comparatively sm^ll 
dispersion best contribute to these investigations by supple- 
menting daily researches? The chromospheric soectra ?*re 
investigated in order to ascertain: (i) wave-lengths; (2) in- 
tensities of the lines; and (3) the heights in miles or kilo- 
meters that the vapors extend above the photosphere. 

It will be shown later that the intensities of the lines of the 
flash spectrum depend on a great many different considera- 
tions, the timing of the exposures, the heights and whether 
the spectrograph is with slit or slitless. For refined work 


where the intensities are measured by a microphotometer, all 
of the various factors must be taken into account. If, how- 
ever, the intensities are estimated on an arbitrary scale in 
the manner of Rowland's Tables, these estimates may be 
carried out with about equal facility no matter what the 
type of instrument. Consequently, in making a decision 
regarding the best type of spectrograph, the question of in- 
tensities may be laid to one side. One has a choice that 
is accordingly much limited. Are exact wave-lengths of the 
chromosphere of the highest importance? 

If the decision is made in the affirmative, then it might 
seem that we should adopt a slit spectrograph. With equal 
dispersion and equally sharp focus, the wave-lengths derived 
from a slit spectrograph will however be superior in accuracy 
to those with slitless instruments in the measurement of the 
strongest lines of the spectrum only. If the observer is care- 
ful to place the slitless spectrograph in such a position that the 
tangents at the centers of the chromospheric arcs lie parallel 
to the lines of the grating or to the edge of the prism, or in 
other words, so that the spectral lines are perpendicular to 
the length of the spectrum, then these spectra are sharp and 
clear-cut and wave-lengths can be determined with high pre- 
cision. With the strongest lines in all spectra there is always 
a spreading of light on the photographic plate by irradiation. 
If a slit is employed this spreading is fairly symmetrical, 
while without a slit the spreading is not symmetrical due to 
the interposition of the moon and to the elevation of the 
heated vapors above the surface of the photosphere. Slitless 
spectra of the chromosphere consequently demand greater 
experience and judgment in the person who measures these 
spectra with the result that the precision of the wave-lengths 
will depend to a large measure on the skill of the measurer 
in his ability to allow for the effects of irradiation. 

With the small amount of light available at eclipses and 
brief exposures of a few seconds, it is possible only to utilize 
a dispersion small in size when compared with the much 
greater dispersion in the every-day solar investigations, such 
as are carried out with the i5q-foot tower telescope and 75- 
foot spectrograph of the Mount Wilson Observatory. A 


dispersion of about 2 or 3 angstroms to the millimeter is 
about the maximum possible with the flash spectrum and this 
permits an accuracy of wave-lengths of about 0.02 angstroms. 
Such an accuracy is much inferior to that employed in ordi- 
nary solar work, and is not sufficiently high to permit the 
determination of systematic differences between eclipse 
wave-lengths and those taken under ordinary solar condi- 
tions. Accordingly, chromospheric wave-lengths practi- 
cally can serve no other purpose than the identification of 
lines for accurate comparisons with Rowland's Tables in 
order to determine the sources whence the spectral lines 

As a matter of fact, no photographs of the flash spectrum 
taken with a slit have furnished any more accurate wave- 
lengths than those of the 1905 eclipse taken by concave 
grating without slit. It therefore appears open to question 
whether spectrographs with slits will give wave-lengths of 
greater accuracy than those of the slitless variety. 

If, therefore, we assume that both wave-lengths and in- 
tensities of the spectral lines may be determined with equal 
accuracy with slit and slitless instruments of about the same 
dispersion, the choice of the best type of spectrograph for 
the flash spectrum must be decided by the ability of the 
two different types of instruments to give information re- 
garding the heights to which the chromospheric vapors extend 
above the surface of the photosphere. The decision can 
now be made in the easiest possible manner. Spectrographs 
without slits give vastly more information about levels than 
can possibly be derived from any form of slit instrument. 

It is, accordingly, the firm opinion of the writer, who has 
confined his attention at eclipses exclusively to the photog- 
raphy of the flash spectrum, that an observer who chooses a 
slit spectrograph to photograph the flash spectrum delib- 
erately thrusts to one side the most valuable information that 
can come from the chromospheric spectrum, namely, the 
heights to which the vapors extend above the sun's surface. 
On the other hand, it is well not to forget that a spectrograph 
with slit readily permits a more accurate and reliable de- 
termination of good focus under the temporary conditions of 


eclipse observation than is possible with any form of slitless 
instrument. It would therefore perhaps be good advice to 
say that if an astronomer is to observe his first eclipse and 
wishes to do some work of value, he had better try to photo- 
graph the flash spectrum using a slit spectrograph with fair 
dispersion, for instance, about equal to that of the Mills 
spectrograph of the Lick Observatory or the Bruce of the 
Yerkes Observatory. For such work it will be well for him to 
confine his attention to the blue end of the spectrum on ac- 
count of the greater dispersion of his spectrograph in this 
region and on account of the greater number of lines in the 
flash spectrum. 

It is impossible to exaggerate the importance of securing 
the photographs of the flash spectrum at the proper instants 
of time so as to secure the spectra of the layers of the chromo- 
sphere as close to the photosphere as possible. At an eclipse, 
there are two manifestations of the flash, one at the begin- 
ning and one at the end of totality. Before the begin- 
ning of the total eclipse, the Fraunhofer lines persist as long 
as there is any portion of the photosphere visible, but when 
the moon entirely covers the sun's surface, or at the very in- 
stant of the beginning of totality, there is the sudden re- 
versal of the Fraunhofer spectrum to that of bright lines. 
If one watches the phenomenon visually with some form of 
spectroscope, he will see many of the high level lines reversed 
many minutes before totality, particularly at the cusps. 

The plan for securing photographs at the proper times al- 
ways followed by the author is an old familiar one. A pair 
of old-fashioned binoculars is used and over one lens there is 
a direct-vision spectroscope. The one used in recent eclipses 
is a replica transmission grating of 15,000 lines per inch. 
With this it is possible to observe any particular portion of 
the spectrum one wishes. With a pair of binoculars and such 
an attachment, it is possible with the left eye to watch the 
disappearing crescent of the sun, shielding the eye by smoked 
or colored glass, while with the right eye the emission lines 
can be watched as they appear one after the other with the 
approach of second contact. Armed with this, the first flash 
can be observed and the exposure started with great nicety. 


On account of the uncertainty in our knowledge of the exact 
duration of totality still persisting, one will be wise to de- 
cide that the exposure for the second flash should begin five 
seconds before the calculated end of totality, and should ter- 
minate with the first trace of the reappearing sun. A delay 
of one-tenth of a second in ending the exposure may readily 
bring ruin to this exposure. In photographing the flash spec- 
trum at an eclipse it is evident that the important photo- 
graphs are two, one at the beginning and one at the end of 
totality. Ordinarily, additional photographs are made, just 
before and immediately after totality, for the Fraunhofer 
and any emission lines. During totality, several short ex- 
posures are given, just after the first flash and again before 
the second flash, for the vapors of greater elevation, with a 
long exposure or two at mid-totality to obtain the spectrum 
of the corona which appears as a series of complete rings. 

My little pair of binoculars just described gave me at the 
1930 eclipse a most gratifying view of the flash spectrum, 
both at the beginning and at the end of totality. One great 
handicap in eclipse work is that in one's lifetime there are so 
few opportunities to rehearse. The 1930 photographs show 
that my reaction time (like that of everybody else who 
has ever photographed the flash spectrum ) is not instantane- 
ous; the first flash does not record the chromospheric lines 
of lowest level. At the end of totality, I kept the exposure 
going as long as I dared, and the photograph shows almost 
perfect timing. 

It need hardly be added that the times of first and second 
flash, recorded preferably on the chronograph, will furnish 
excellent observations of the beginning and ending of the 
total eclipse. On account of the rough character of the 
moon's edge, these times will not agree with those recorded 
from the last appearance and first reappearance of Baily's 
beads which brings up anew the time-honored question of 
what one means by the " beginning " and " ending " of a 
total eclipse. 

The above descriptions of slitless instruments and meth- 
ods of obtaining the flash spectrum all refer to a stationary 
photographic plate where the lines of the spectrum consist of 



a series of crescents. At the eclipse of 1898 in India, Camp- 
bell tried out an ingenious innovation of placing a narrow 
slot directly in front of the photographic plate in the manner 
shown in Figure 5 ; and then moving the photographic plate 
uniformly in a direction perpendicular to the slot. The same 
procedure was followed at the eclipse of 1900 in Georgia, 
1905 in Spain and 1908 in Flint Island. The best definition 
of the four eclipses was obtained in 1905. After a delay of 

FIG. 5 The moving-plate spectrograph 

more than twenty years, a discussion by Menzel, together 
with an introduction by Campbell, has recently appeared in 
handsome form as Volume XVII of the Publications of the 
Lick Observatory. This volume forms a magnificent con- 
tribution to the study of solar problems, quite in keeping with 
the high standard of excellence set in the past forty years in 
the observation of eclipses by the Lick Observatory and by 
Dr. Campbell, its director for much of this period. The 
author wishes to express his highest praise of the beautiful 
Lick spectra and of the splendid character of the discussion. 
If the moving plate is set in motion about twenty seconds 
before the beginning of totality, it will automatically record 
the gradual appearance as bright lines, first of the lines of 
highest level, then of medium level lines and then the spec- 
tacular appearance of the flash spectrum in the manner 
described by Young at the eclipse of 1870. At the end of 
totality the plate should be started more than five seconds 
before the calculated end of totality in order to photograph 
the second appearance of the flash spectrum in reversed or- 

< B 

& s 


o _ o 

2 3-jj 

r<- ON "" 

fc c 


W .00 o 
" tfl 1> 

I 5 

S 2 w 


p d ^ 

OJ ' '"' 


j 73 



Fi 4) Jj, 

P a 



W & 

2 * 



der. Facing page 260, there is a reproduction of the Lick 
moving plate of the second flash of the 1905 eclipse. The 
strip of normal Fraunhofer spectrum of the sun appended 
below for comparison was taken by Menzel with the original 
1905 apparatus reassembled after more than twenty years 
and with the addition of a slit and collimator lens. The 
streaks running lengthwise throughout the spectrum are 
caused by the non-uniform motion of the photographic plate. 

It is interesting to compare the details of this moving plate 
spectrum with that taken by the author with fixed plate, as 
shown facing page 248. The spectra taken by the two 
methods refer to the same (1905) eclipse and both were 
taken at the end of totality. In the original spectra, in the 
region near the strong H and K lines shown to the left on 
page 248, the Lick moving plate had three times the disper- 
sion of the fixed Mitchell plate. 

One obvious and important advantage of the moving over 
the fixed plate is that the former makes less drastic demands 
on the ability of the observer to time his exposures so as to 
reach the lowest possible levels. Another advantage is that 
the moving plate records the gradual changes from the dark- 
to the bright-line spectrum at first flash, and in the reversed 
order at second flash. The fixed plate makes the record in 
discontinuous exposures, the moving plate has a continuous 

Now that the Lick results have been published, it is pos- 
sible to compare the relative merits of the fixed and moving 
plates for the photography of the flash spectrum. When 
the performances of any two spectrographs are compared, it 
is almost the universal custom among skilled spectroscopists 
to draw conclusions only when the two instruments have 
nearly identical optical powers, but more particularly when 
the dispersion is nearly the same for each. By disregarding 
this basis of comparison, many mistakes have been made in 
interpreting the results of the flash spectrum, the most out- 
standing probably being that by Lockyer, referred to on page 
162, who compared the number of lines in the flash spectra 
photographed in 1893 an d 1896 with the number found in the 
great Rowland Atlas obtained with vastly superior definition 


and dispersion. One is therefore surprised to find Campbell 
(loc. cit.) comparing the number of spectral lines of the 
1905 and 1908 moving plates with those given by the 1905 
fixed plate with concave grating, and then drawing conclu- 
sions in spite of the fact that the 1905 moving plate had twice 
the dispersion, and the 1908 plate four times the dispersion 
in the region of wave-lengths common with the concave grat- 
ing photograph. 

Investigations of the flash spectrum are attempted in order 
to photograph the solar spectrum under conditions differing 
radically from those of every-day observation without an 
eclipse. The heights to which chromospheric vapors extend 
are of the very greatest value in furthering solar research. 
Hence, if comparisons are made of fixed and moving plate 
spectrographs having the same dispersion, then it should be 
easy to decide which method will furnish heights with the 
maximum of accuracy and which will be most efficient in re- 
cording the greatest number of lines. Of course, it goes 
without saying that the dispersion of both fixed and moving 
plate spectrographs should be as great as possible. A com- 
parison is not now being made of the relative merits of prism 
or grating spectrographs for the reason that either type of 
instrument may be combined with either fixed or moving 

Under ideal conditions of perfect timing and circular edges 
of both moon and sun (the moon without mountains and val- 
leys, the sun without prominences), the flash spectrum ap- 
pears at the instant that the whole photosphere is covered up 
by the moon, hence with edges of sun and moon tangent as 
in Figure 6. The fixed plate records the crescent arcs. A 
knowledge of the angular diameter of the sun and the aug- 
mented diameter of the moon furnishes a ready method of 
deriving from the spectra the heights to which the chromo- 
sphere vapors extend. Measurements are made of the angu- 
lar length of the different arcs, or more simply by the meas- 
urement of the half -chord of this arc yo in Figure 6. Under 
the ideal conditions outlined above, the chromospheric arcs at 
their middle, or at the horizontal line in Figure 6, represent 
in each spectral line the radiations from all levels of the 


chromosphere lying above the photosphere. At a height of 
y<> equal to 2.0 mm on the original spectra of 1905 (with a 
focal length of 60 inches) each of the crescent arcs repre- 
sents the chromosphere at levels that are above 700 kilo- 
meters, a half-chord of 3.0 mm in each case gives the condi- 
tions at levels above 1600 km. At a value of y<> equal to 5.0 

FIG. 6 Measuring the heights of solar vapors. 

mm, the tip of a high level cusp is reached and this tip fur- 
nishes a detectable amount of radiation impressed on the 
photographic plate. At the tip of this chromospheric arc, 
therefore, only those radiations reach the photographic plate 
which come from emitting atoms above the 5000 km level. 
Hence each chromospheric arc, both above and below the 
middle of the arcs as recorded in the spectrum, gives a record 
of the radiations at the photosphere and at higher and still 
higher levels as the tips of the arcs are approached. If the 
tip of the chromospheric arc comes at a half-chord measured 
at 2.0 mm, then the radiations above the 700 km level are 
sufficiently active to give a blackening of the photographic 
image strong enough to be detected. The " height " is then 
said to be 700 kilometers. Under the actual conditions at 


the time of the eclipse, neither the edge of the moon nor that 
of the sun is circular in outline, nor is the exposure of the 
flash spectrum an instantaneous one, and the scale of heights 
is accordingly modified. The moon has mountains and val- 
leys projected against the edge of the sun, which has promi- 
nences. The chromospheric arcs will present a picture of 
these effects. If these arcs are very irregular, then mani- 
festly it would be difficult to get reliable information re- 
garding heights, particularly if a large and active prominence 
is projected near the tips of the chromospheric arcs. 

With the moving plate method, all features of the chromo- 
spheric arc are exactly as for the fixed plate outline above, 
the only difference being that the slot used with moving plate 
concentrates attention on a definite section of the chromo- 
spheric arc. With the arrangement used at the 1905 eclipse, 
the slot width uncovered a layer of 300 km at one time. As 
the eclipse progressed the moving plate furnished an inte- 
grated picture of what was happening within this layer 300 
km in height. 

In the earlier editions of this book, and also in the memoir 
in Volume IV of Handbuch der Astrophysik, the author has 
stated that one decided objection to the moving plate method 
is that at the time of an eclipse the slot might readily be 
superposed on a section of the sun where there was an erup- 
tive prominence. At the eclipse of 1918, the " Heliosaurus " 
shown facing page 196 was exactly at the position of third 
contact. At the eclipse of 1919, another large prominence 
group is shown facing page 272. 

The Lick observers recognize the force of this criticism as 
is shown by the following (loc. cit. p. 249), " One valid ob- 
jection to the moving plate is that, if the image of a promi- 
nence falls on the plate, the resultant heights would be diffi- 
cult to interpret." Hence at the latest eclipse observed with 
moving plate, that of April 28, 1930, the Lick observers ex- 
amined the solar image just before the beginning of totality 
in an effort to set the slot on a portion of the sun devoid of 
large disturbances. One unfortunate feature resulting from 
this attempt * was the failure to photograph the layers of 

1 Publ. A. S. P., 42, 131, 1930. 


lowest level owing to the plate not being put in motion until 
after the total phase was well started. 

Regarding the relative number of lines photographed by 
fixed or moving plates, it seems easy to reach conclusions. In 
the region of wave-lengths, from 4200 to 4500 A, where all 
three spectra had the best definition, the 1905 and 1908 mov- 
ing plates recorded 425 and 292 lines respectively, while in 
the same region of wave-lengths, the 1905 concave grating 
spectra had 43 7 lines ; and this in spite of the fact that the 
moving plates had average dispersions twice and four times, 
respectively, that of this 1905 fixed plate. Between the re- 
gion 3900 and 4100 A, the Lick 1905 moving plate recorded 
413 lines and the Lick 1905 fixed plate 454 lines in spite of 
the better definition of the former and a dispersion more 
than twice greater than that of the latter. Hence with equal 
dispersion and definition the fixed plate will always show 
more lines than a moving plate. 

Another disadvantage of the moving plate is the difficulty 
of ascertaining what zero to adopt as the initial level for the 
scale of heights. This zero-point depends on the reversal 
from dark to bright lines on the photograph (at first flash) 
or the reverse at second flash. As already stated, at the be- 
ginning of totality, the high-level lines become bright many 
seconds before those of lower levels. The reversal of the 
lines of lowest level forms the " flash " spectrum. With 
the fixed plate, measures of the lengths of the arcs give the 
heights directly, but with the moving plate, the photographs 
themselves must furnish the zero-point from the reversal of 
the lines. This is a difficult task which is further compli- 
cated by the fact that the slot causes an effect integrated 
over a region stretching over 300 km in height. It would 
seem therefore that this integrated effect would not permit a 
high precision in the determination of heights by the mov- 
ing plate which must also be subject to a zero-point correc- 
tion far larger than with the fixed plate. 

In making comparisons of the reliability of fixed and mov- 
ing plates, the best conclusion can be reached by confining 
attention to the low-level lines only, those that reach heights 
of 700 km or less, these lines being vastly greater in number 


than those of higher levels and their heights are known with 
greater precision. The Lick 1905 fixed plate gave average 
heights for these lines of lowest levels which differed about 
50 km from those of the 1905 Mitchell plate. On the other 
hand, the average heights from these same low-level lines 
from the 1905 moving plate were 400 km greater than those 
from the Lick 1905 fixed plate. At one and the same eclipse, 
the simplest method of interpreting this systematic differ- 
ence is the zero-point correction of the moving plate. More- 
over, the 1905 moving plate gave average heights for the low 
levels more than 500 km greater than the 1908 moving 

Furthermore, with the moving plate, the scale of heights 
depends on an accurate knowledge of the speed of motion 
of the plate and also on the assumption that this motion is 
absolutely uniform. 

Considering therefore the difficulties outlined above and 
the various causes of errors that may beset the moving plate 
method, the author is in hearty agreement with the opinion 
expressed (loc. cit., p. 2) that " it is doubtful if any other 
astrophysical instrument is as difficult to set up and operate 
successfully as an objective-prism eclipse spectrograph, when 
used with a continuously moving photographic plate includ- 
ing a slit immediately in front of the plate." One might also 
add that it is doubtful if the photographs of any other astro- 
physical instrument are so difficult to interpret in order to 
reach conclusion free from systematic errors. 

If the flash spectrum were an exact reversal of the Fraun- 
hofer lines, both as regards wave-lengths and intensities, 
the eclipse observations could add but little to our knowl- 
edge of solar physics. In such a case, the precious moments 
of a total eclipse should be devoted to the investigation of 
other lines of research. In fact, the flash spectrum is inter- 
esting and important only in so far as it differs from spectra 
taken under ordinary conditions. 

The intensity of a spectrum line depends both on the width 
and on the blackness of the photographic image of the line. 
It is unfortunate that in all spectra, whether of dark or bright 
lines, whether determined in the laboratory or in the observa- 


tory, it is ordinarily impossible to have a scale for the desig- 
nation of relative intensities which is other than arbitrary. 
With such a scale the strongest line in any spectrum may be 
represented by 10, by 100, or even by 1000, while the weak- 
est line receives the number i, or o, i, 2, or even 
3, as in the revised Rowland's Tables. Such scales being 
arbitrary may not be uniform, and it is consequently very 
difficult to compare the values of the intensities of one spec- 
trum assigned by one observer with those investigated by 

For reasons outlined above, with chromospheric arcs pho- 
tographed by the fixed plate, it is readily possible to make 
estimates of intensities or, better still, measures of intensities 
by the microphotometer at different positions along the arcs. 
At the middle of the arcs the intensities refer to the radiations 
of all atoms above the photosphere. At positions 2 mm from 
the middles of the arcs the intensities are thus derived from 
the radiation from all atoms above 700 km in height. By 
estimating or measuring the intensities at different places 
along the chromospheric arcs, from their middle to the tips of 
the cusps, a knowledge is obtained of the radiations of atoms 
at higher and still higher levels above the photosphere. In 
similar manner with the moving plate, estimates of inten- 
sities may be made or measures with the microphotometer 
be carried out, as was done by Menzel (loc. cit.}. With the 
moving plate, as with the fixed plate, the intensities, esti- 
mated or measured, refer always to the radiations from atoms 
above certain levels as determined by the edge of the ad- 
vancing moon. 

As has been stated, with the fixed plate, the interpreta- 
tion of heights is complicated by the Baily's beads which 
make knots in the curved lines of the spectrum. However, in 
estimating or measuring intensities on the fixed plate at dif- 
ferent heights above the zero-point, the observer has it quite 
in his power to make measures at any particular location 
along the whole of the curved arc. Hence the Baily beads 
need be of no particular concern unless the solar disturb- 
ance is great in size and extent. With the moving plate, on 
the contrary, the slot has a fixed definite position and meas- 


ures may be made of conditions affecting the chromosphere 
at this and at no other location. 

The most characteristic difference between the chromo- 
spheric and the Fraunhofer spectra is found in the relative 
intensities of the lines. The system of intensities adopted 
for the chromospheric spectrum is purely an arbitrary one, 
in which 200 represents the strongest lines like K and Hy, 
and o that of the weakest line. Naturally the intensities 
depend on the character of the photographic plate used, but 
partial allowance may be made for the decrease in sensitive- 
ness of the plate in the green and yellow regions. In estimat- 
ing intensities, one is unconsciously influenced by the breadth 
of the lines, so that the values for intensity give a somewhat 
combined appraisal of the blackness and breadth of a cer- 
tain line. These at best are but estimates, but they are 
perhaps comparable in accuracy with estimates of intensities 
by others, for instance, as in the Rowland's Tables. Esti- 
mates in Rowland and in the flash spectrum may 1 be cali- 

The reasons for these pronounced differences in intensi- 
ties between the dark line Fraunhofer spectrum and the 
chromospheric spectrum will be evident on a moment's re- 
flection. Consider a Fraunhofer line coming from the center 
of the sun's surface, and assume that the absorption is caused 
by a reversing layer 500 miles in thickness. The light com- 
ing perpendicularly from the photosphere can be absorbed 
by the atoms in this 5oo-mile layer. At the time of an eclipse, 
the light of the chromosphere comes tangentially from the 
sun's surface, and not perpendicularly, with the result that 
the chromospheric light is affected by a depth of 20,000 miles 
of atoms. (The line of sight from a layer 500 miles in thick- 
ness passes through 20,000 miles of solar atmosphere when 
tangent to the sun's surface.) This has an important bearing 
on Saha's theory, as will be explained in Chapter XVII. 

Researches on the structure of the atom, as shown in 
Chapter XIV have demonstrated the important fact that the 
intensity of a line in any spectrum depends primarily on the 
number of emitting or absorbing atoms. In the Fraunhofer 

1 Astro physical Journal, 68, i, 1928. 


u -d 



(S ** 







spectrum the atoms are absorbing radiation but in the 
chromospheric spectrum they are emitting radiation. Let us 
now consider two different elements in the sun's envelope; 
one of these elements has a low density but extends high in 
miles above the photosphere; the other element is much 
heavier and its atoms are confined to layers lying much closer 
to the sun's surface. It is easy to imagine in the case of ab- 
sorption of radiation that the total number of atoms involved 
in producing a line in the spectrum of the lighter vapor might 
be identical to the number contributing to the dark line in 
the heavier vapor on the condition where the light passes 
radially through the gases, for instance, when coming from 
the center of the sun's disk. If , however, at the time of a total 
eclipse when the moon covers the photosphere, the atoms of 
the chromosphere emitting radiation come tangentially from 
the sun's envelope. Under these circumstances, the greater 
height attained by the atoms by the rarer of the two vapors 
adds enormously to the number of atoms encountered rela- 
tive to the number encountered in the low-lying and denser 
layer when the atoms are excited and radiate light. Hence, it 
is readily seen that although the two gases may give lines of 
equal intensity in their absorption spectra, they will not 
necessarily do so in their emission spectra; the low-lying 
heavy vapor will give in the chromospheric spectrum short 
arcs, while the other assumed vapor will give longer arcs of 
greater relative intensity. Though there are many contribut- 
ing causes, the main reason for the great differences in in- 
tensities between the dark and bright line spectrum of the 
sun is the differences in heights to which the vapors extend. 
The H and K lines of enhanced calcium and the lines of the 
hydrogen series are the strongest lines in the chromosphere 
mainly for the reason that the atoms of enhanced calcium and 
hydrogen are detected at greater elevations than are attained 
with any of the other elements. 

As a matter of fact, there are such enormous differences 
in the intensities of the Fraunhofer and flash spectra that 
placed side by side, as they are on page 268, the spectra 
seem to belong to stars of two different types rather than 
to the same object under different conditions. It is these 


differences that make observations of eclipse spectra of the 
greatest value in widening our knowledge of solar physics. 
The chief differences in intensity for the stronger lines are 
found in the elements helium and hydrogen. As is well 
known, no helium absorption lines are ordinarily found in 
the sun, whereas in the eclipse spectrum the helium lines 
are conspicuous by their great strength. In the Fraunhofer 
spectrum there are only four hydrogen lines visible, while in 
the flash spectrum there is the whole Balmer series, no less 
than thirty- four lines being measured on the 1905 plates. 

One can therefore see at a glance the very great importance 
attaching to a knowledge of heights in the chromosphere. 
The following chapters will show how this information not 
only has been the basis of Saha's Theory of lonization but 
in addition has had an important practical bearing on many 
solar and stellar problems. 

This close comparison of the solar and chromospheric spec- 
tra side by side helped make the identifications of sources 
rather certain. But the Rowland spectrum had a dispersion 
fully ten times the dispersion of the chromospheric spectrum 
(2i-foot radius in the second order compared with 5-foot 
focus in first order, the gratings having nearly the same num- 
ber of lines per inch). Naturally, lines which appear single 
in the chromospheric spectrum may be a blend of two or 
more lines with the greater dispersion. But lines which ap- 
pear as a close pair or a blend in the chromosphere must be 
the result of the blend of corresponding lines in Rowland. 
On account of the great differences in intensity of the chro- 
mospheric and Rowland spectra, it was difficult to be always 
sure of identifications until photographs were compared side 
by side. The original photograph of the flash spectrum was 
enlarged five times. Rowland's great Atlas was reduced six 
times. Since the flash spectrum was nearly normal, it was 
possible to procure both spectra on a comparatively close 
approximation to scale. This comparison of spectra will 
perhaps speak more strongly, than any words or comparison 
of wave-lengths, concerning the sharpness of the original 
spectrum of the chromosphere. On account of the small 
variations from the normal spectrum (noted above) it was 


impossible to obtain an exact match in scale in the two pho- 
tographs. Those who are interested sufficiently will be able 
to carry the comparison along line for line. 

Repeated attention must be directed to the fact that the 
heights of the chromospheric vapors determined from the 
measurement of the angular length of the cusps, can afford 
no great accuracy in the determination of the absolute heights 
in kilometers to which the various layers extend above the 
photosphere. The method depends on the visibility of the 
ends of the cusps. It is quite possible, and probable, that 
vapors extend in detectable amounts to elevations beyond 
the limits visible in the cusps. The heights derived by this 
method can therefore only represent a mean height and can- 
not be expected to furnish the maximum heights to which 
the vapors in detectable amounts extend. Attention should 
likewise be called to the fact that the heights measured in 
this manner cannot give the elevations above the photo- 
sphere, but rather above the average level of the layer pho- 
tographed in this particular flash spectrum. With these limi- 
tations, the method is capable of furnishing the relative 
heights of the layers producing the individual spectrum lines 
with a fair degree of accuracy. 

This is not the place to go into greater technical details 
which may be found in Publications of the Lcander McCor- 
mick Observatory, Vol. 5, part 7, 1932, or in the Astrophysi- 
cal Journal. However, a brief summary may be made here 
of the advantages and disadvantages of the moving over the 
fixed plate for the spectrum of the chromosphere. It is as- 
sumed that the two types of spectrographs have about equal 
dispersion and light gathering power. Wave-lengths are ob- 
tained with equal accuracy by the two methods and hence 
may be left out of consideration. 

The moving plate is a much more complicated instrument 
and demands greater technical skill in the erection and ad- 
justment at the eclipse site. In photographing the flash 
spectrum, the moving plate has distinct advantages in the 
proper timing of the exposures. The moving plate shows the 
gradual change from dark to bright lines at the first flash (or 
the reverse on second flash). The separate exposures with 


the fixed plate make a similar but less complete record. With 
equal dispersion and definition, the fixed plate records more 
lines than the moving plate. 

If the chromosphere is greatly disturbed near the posi- 
tions of second and third contacts, the heights recorded from 
the lengths of the arcs on the fixed plates are difficult to inter- 
pret, especially if a solar prominence occurs near one of the 
cusps. With the moving plate, the slot which defines the 
region of investigation may be superposed on an active promi- 
nence region and therefore the photographs will not repre- 
sent average solar conditions. For the first flash, the slot 
might possibly be set immediately before totality so as to 
avoid disturbed regions in the chromosphere. This however 
could not be done for second flash. The present writer would 
never under any circumstances attempt such an operation. 
It is far too difficult a task to undertake in the few excited 
seconds available. It is far easier to observe the appearance 
of the flash spectrum with the binoculars and direct-vision 
spectroscope, and then to start and close the exposures ac- 

If successful photographs are obtained with both fixed and 
moving plates, it is much easier to interpret the results of the 
former. With the arcs recorded by the fixed plate, any por- 
tion of them may be adopted for discussion. The levels that 
lie closest to the photosphere manifest themselves by the 
appearance of a strong continuous spectrum on the photo- 
graph. Hence with the fixed plate, the zero-point correction 
should be small, probably of the order of 50 kilometers. At 
the 1905 eclipse, the Lick fixed plate at second contact and 
the Mitchell fixed plate at third contact gave low-lying levels 
which differed about 50 km, those of the latter recording the 
greater heights. 

With the moving plate, the zero-point correction depends 
on the estimation of the location where on the photograph the 
spectral lines reverse from dark to bright (or bright to dark). 
The zero-level should be that of the vapors closest to the 
photosphere, those of lowest level. If there are many en- 
hanced lines in the spectrum, as in the region of 4300 and 
4400 A, the problem becomes complicated and more diffi- 

Photographed at Yerkes Observatory by Pettit with the spectroheliograph 

nn tnr> Ha\r nf fl-i f/^tol ^K^o. T>U -L__.L i_^- i , ,. . . . ^e>*c*i-'*i 

on the day of the total eclipse. 

The greatest height attained on that day was 
760,000 kms. 


The same prominence is shown facing page 272 from photographs taken at Yerkes 
Observatory where there was no eclipse, 


cult. As already stated, the Lick 1905 moving plate gave 
the heights of the vapors of lowest level 400 km in excess of 
the Lick 1905 fixed plate. This systematic difference indi- 
cates zero-point corrections which are larger for moving than 
for fixed plates. 

The 1905 Lick moving plate had a slot width of .05 inches 
which permitted radiation from 300 kms of vapors to pass 
through and be recorded on the photograph. The motion of 
the plate causes progressive exposures to be made along the 
spectral line. The height of the vapor is derived in each case 
from the total length of the line measured from the assumed 
zero-point up to the vanishing tip. But this total length 
depends on the integrated effect of the radiation passing 
through the slot of the mechanism. Hence it would seem 
that the moving plate might readily have systematic differ- 
ences between the low-level heights and those of greater ele- 
vation, and this is further complicated by departures of the 
motion of the plate from a uniform character. " An examina- 
tion of the 1905 eclipse plate indicates that in the neighbor- 
hood of 1800 kms, the plate had experienced a slight accelera- 
tion " (loc. cit., p. 258). 

In regard to the photometric measures, a quotation may 
be made (in its corrected form) by Campbell (loc. cit., p. 
iv). " My moving-plate method gives the average or in- 
tegrated effect of a relatively long segment of the chromo- 
spheric crescent itself, say 300 km long, the inner edge of 
which is serrated by lower mountain peaks. The fixed plate 
method records the serrated crescent as a whole (Mitchell), 
or in part (Pannekoek and Minnaert) ; and it is interesting 
to note, a long segment of the serrated photographic crescent 
was averaged or integrated upon a moving plate of the re- 
cording microphotometer by Pannekoek and Minnaert as 
the basis of their photometry of the chromospheric spectrum. 
In their photometric applications, therefore, the two meth- 
ods seem to have an element of equivalence." However, we 
might remind Dr. Campbell that in the measurement of the 
fixed plate spectra, it is not necessary to use a long slit with 
the microphotometer. A short slit may be used in order to 
integrate the measures over a relatively short range in heights. 


In the actual operation of measuring the intensities of the 
spectra of fixed or moving plates there should be practically 
no difference in procedure. 

In the foregoing, there has been stressed some few of the 
difficulties besetting the eclipse observer. No matter how 
carefully and precisely the focus of the spectrograph is de- 
termined, no matter how accurately the exposures are timed, 
either by fixed or moving plate, the photographed spectra 
may not be of finest quality for causes entirely outside the 
control of the observer, namely, the " seeing." As an ex- 
ample, the experience of the writer may be cited. At the 
1925 eclipse the flash spectrum was photographed with the 
same equipment as at the 1905 eclipse. At the 1925 eclipse, 
the sun was only 17 above the horizon, and in consequence 
the seeing was poor. The 1905 eclipse was observed under 
excellent conditions of seeing. The 1905 spectra show the 
faintest lines of the spectrum in much greater numbers than 
the 1925 photographs, the lines in the former eclipse being 
stronger and more clearly defined. The explanations for the 
differences in the two spectra cannot be found in differences 
of focus, or of exposures, or of levels photographed. Under 
the poor conditions of seeing of the 1925 eclipse, the light of 
the bright-lined chromospheric arcs was rendered ill-defined 
by being spread over a larger area. With the strong lines of 
the spectrum, this effect made the edges hazy but diminished 
the total intensities of the lines but little. With the weak 
lines, however, the spreading of the light over a larger area on 
the plate caused a marked diminution in intensity and sharp- 
ness of each line with the necessary result that the weakest 
lines in the 1925 spectrum were practically obliterated by 
the poor seeing and only those lines of a certain minimum 
strength survived. Astronomers who are engaged in the ob- 
servational work of stellar photography, particularly with 
telescopes of great focal length, are entirely familiar with an 
exactly analogous problem. Under the best conditions of 
seeing, the star images are hard and sharp and with well- 
defined edges. Under poorer and poorer conditions of see- 
ing, the star images increase in size and the edges become 
more and more fuzzy. The larger stellar images with poor 


seeing require an increase of exposure to photograph a star 
of any given magnitude. The length of exposure becomes 
greater and greater as the seeing deteriorates more and more. 
Under poor conditions, the star images on the plates are 
large and ill-defined and the accuracy of measurement, when 
compared with plates taken under good conditions, is much 

If photographs of the flash spectrum are actually secured 
in spite of all difficulties to be overcome, then comes the 
much more important problem of the discussion. The size 
of this task may be surmised from the fact that Dr. Menzel 
spent about five years in preparing for publication the hand- 
some Lick Publication, Volume 17. On page 2, it took one- 
half a quarto page for him to render thanks to those individu- 
als or bodies who had given scientific or financial assistance. 
In similar manner it has also taken the present writer, al- 
though working almost alone and with his time interrupted 
by many duties, five years to prepare the original contribu- 
tion in Publications of the Leander McCormick Observa- 
tory, Volume II, parts 2 and 3 and the revision in Volume V, 
parts 2, 3, 6 and 7. 

It will not be necessary here to explain in further detail 
the methods followed in deriving wave-lengths and heights. 
This information may be obtained from the sources just 

When the spectra of the 1905 eclipse were being discussed 
by the writer, comparatively little was known of the differ- 
ences between the conditions existing in the Fraunhofer 
spectrum and in the chromosphere. The main difference 
known to exist was that pointed out by Fowler from the 1898 
eclipse spectra, namely, that the enhanced lines played an 
important role in the flash spectrum by showing increased 
intensities over those in the solar spectrum. But how was 
one to find the wave-lengths of the lines which had greater 
strengths in the spark than in the arc spectrum? Lockyer 
had given lists of some of these lines, and these were good as 
far as they went. But the known lines were few in number 
and entirely inadequate for a discussion of the problem. 
What was to be done? Kayser's important lists had not yet 


appeared. There was nothing to do but to follow the ex- 
ample set by Kayser and others, namely, that of going to the 
original sources one by one. There seemed no other method, 
and this meant many long and tedious hours of patient com- 

Within the past decade, however, the combined attack on 
the structure of the atom by the astronomer, physicist and 
chemist has resulted in a stupendous increase in knowledge 
of conditions underlying the production of emission and ab- 
sorption lines in different spectra. The outstanding achieve- 
ments have been the Bohr-Sommerfeld atom, Saha's theory 
of ionization, investigations of series in spectra coming mainly 
from the hands of Fowler, Russell and Meggers, and the 
publication from Mount Wilson Observatory of the Revision 
of Rowland's Table of Solar Wave-Lengths. As a result of 
this greatly increased information now available, particularly 
in the proper identifications of the sources causing the spec- 
tral lines, the discussion of the flash spectrum has become 
one of greatly increased simplicity. 



A" | AHE depths of the various layers of gases surround- 
I ing the sun have an importance which is very vital 
JL in all theories of solar physics. Eclipse spectra fur- 
nish the only means yet known of directly measuring these 
depths. Comparisons of the flash spectrum with the solar 
spectrum taken under ordinary conditions reveal many im- 
portant conclusions, the first of which is that wave-lengths 
are identical in the flash and in the Fraunhofer spectrum. 
This statement can be true only within the limits of accuracy 
attainable in the measurements of wave-lengths of the flash 
spectrum, or expressed in other terms, it may be said that no 
differences between the two spectra exceed 0.02 angstroms. 
The flash spectrum must therefore be regarded as a true re- 
versal of the ordinary spectrum of the sun since every strong 
line in the solar spectrum, without any exception, is changed 
to a bright line at the beginning and at the ending of totality. 
Although the wave-lengths in the two spectra are identi- 
cal, this is far from being true with the relative intensities 
of the lines. In the flash, many lines appear which are not 
found in the ordinary solar spectrum; furthermore, some 
strong lines in the Fraunhofer spectrum appear as weaker 
lines in the flash, and vice versa, weak lines in the sun are 
strengthened in the spectrum of the chromosphere. Dif- 
ferences in intensity signify differences in level and in elec- 
trical, thermal and pressure conditions of the vapors pro- 
ducing the lines, and consequently an intimate study of such 
dissimilarities will give valuable information regarding the 
distribution of atoms in the solar atmosphere. 

The problem, however, is a complicated one for the reason 
that the intensities depend so intimately on the level of the 
chromosphere photographed, all spectral lines being most 



intense at levels closest to the photosphere. Hence, at one 
and the same eclipse and with exactly the same timing, dif- 
ferent intensities are found on the photographs depending 
on whether the spectrograph is used with or without slit. 
Without a slit, all levels contribute their emission to the 
spectral line, while with a slit only those levels covered by 
the slit give up their light to the line in question. Mani- 
festly, the position of the image of the slit with respect to 
the photosphere is of prime importance, and also whether 
or not there is a prominence at the sun's edge where the slit 
is projected. Hence with a slit, there are many reasons why 
the intensities of the flash lines estimated or measured from 
the photographs of any one eclipse may vary from photo- 
graph to photograph or from instrument to instrument. Even 
on the same photograph there are differences of intensities 
for the same line. In their publication dealing with the 1926 
eclipse, Davidson and Stratton give the estimates at two dif- 
ferent positions in their photographs, called respectively 
" flash " and " prominence." From their photographs taken 
at the 1927 eclipse with slit, Pannekoek and Minnaert note 
similar differences in intensities. 

For the same reasons, eclipse spectra taken without slit 
show wide ranges in intensities depending on the levels pho- 
tographed. For instance, the author secured spectra with 
the same equipment in 1905 and 1925. The intensities dif- 
fer materially in the two years, and in turn they both differ 
in most details from the intensities observed at the 1926 and 
1927 eclipses, and also from the Lick results at the eclipses 
of 1898, 1900, 1905 and 1908. One should, therefore, be 
very guarded when comparing one eclipse with another be- 
fore coming to the conclusion that differences in intensities 
between two eclipses mean differences in solar activity. 
Manifestly one should not push too far any deductions based 
on the intensities at any one eclipse. 

Since the date of the second edition of this book, several 
important discussions have appeared dealing with the ob- 
servations of the flash spectrum. The instrumental equip- 
ment has been described in the preceding chapter. A list of 
the more extended of these publications follows, arranged in 


order of the dates of the eclipses observed. As already stated, 
Campbell with the moving plate obtained photographs in 
1898, 1900, 1905 and 1908. In 1905, with the fine seeing of 
the Spanish eclipse, the spectrum also was photographed on 
a fixed plate. The discussion by Menzel makes a veritable 
storehouse of valuable information. Accurate wave-lengths 
derived from the weighted mean of the different spectra make 
possible a close comparison with Rowland. The identifica- 
tion of the individual lines are given in splendid detail to- 
gether with their multiplet designation. The intensities were 
estimated on the Rowland scale. For certain of the stronger 
lines, the intensities were measured by the microphotometer 
at different heights above the photosphere. On the observa- 
tional side, the most important information is contained in 
the heights from both fixed and moving plates. As shown 
in the preceding chapter, the heights from the moving plates 
are subject to larger systematic errors than those from fixed 
plates. After tabulating the material in order of wave- 
lengths, a further tabulation rearranges the material accord- 
ing to the source of the line, a grouping being made accord- 
ing to the multiplet structure. A summary of the various 
recent theories underlying the problems of solar physics 
makes this Lick publication an epoch-making contribution. 
As an illustration of the tremendous activity since 1923 (the 
date of the first edition of this book) caused by researches 
on the structure of the atom, there may be cited the number 
of references to scientific publications made by Menzel in 
footnotes. Before 1923, references were made to 18 arti- 
cles; in the years 1923-25 there were 22 references; while 
in the interval 1926-30, no less than 100 publications 
appeared, of which number there were 69 citations in 1928 
and 1929. 

Following the eclipse of 1908 observed by Campbell, the 
next eclipse to provide a photograph of the flash spectrum 
was 1914, when Abetti secured spectra of small dispersion. 
At the eclipse of 1918, on account of widespread clouds, the 
only photograph of the flash spectrum in good detail was 
obtained by H. C. Wilson with concave grating, but also with 
small dispersion. No attention was paid to the spectrum of 


the chromosphere in 1919 and 1922, the Einstein problem 
being of paramount importance. Little progress was made 
at the 1923 eclipse. 

At the 1925 eclipse, Curtis and Burns 1 with a concave 
grating of short focal length obtained photographs at the red 
end of the spectrum on plates stained with dicyanine. About 
100 lines were measured between the D lines and 8807 A, 
the dispersion being small, of 80 angstroms per mm. At this 
same eclipse, Anderson with a 21 -foot concave grating with- 
out slit secured spectra on which the definition was vitiated 
by the poor qualities of seeing. At the same eclipse, with 10- 
foot concave grating without slit, Mitchell obtained photo- 
graphs of the flash spectrum. For publication, the results 
were combined with those of the 1905 eclipse photographed 
with the same equipment. At the extreme ultra-violet, the 
measures of the 1926 spectra of Davidson and Stratton were 
included in the discussion by Mitchell. This publication, 
like that of Menzel, involved the newer ideas of the structure 
of the atom. In fact, both Menzel and Mitchell, through the 
kindness of the Mount Wilson astronomers, had free access 
to the valuable compilations on multiplet structure. The 
region of wave-lengths included in Mitchell's discussion was 
from 3066 to 7065 A, while that of Menzel included the re- 
gion 3229 to 5328 A. 

In 1926, Davidson and Stratton " secured beautiful photo- 
graphs with a quartz slit-spectrograph giving excellent defi- 
nition from 3066 to 3968 A. With a camera of i9-feet focus 
and direct-vision prism, the region photographed was ex- 
tended to H(3, and with camera of 4O-feet focus and prism 
of 40, the region was further extended to beyond Ha. The 
photographs with the two prismatic cameras were taken with- 
out slit and furnish heights for the stronger lines of the spec- 
trum. The results were grouped according to multiplets 
shown by the different chemical elements. 

Davidson and Stratton compared the heights derived from 
their prismatic cameras with those of the 1905 eclipse with 
concave grating. Differences were noted which depend 

1 Publications of Allegheny Observatory, 6, 95, 1925. 

2 Memoir R. A. S., 64, 105, 1927. 


mainly on definition but there was no systematic error in 
zero-point between the heights from the two eclipses. 

At the 1927 eclipse, Pannekoek and Minnaert secured ex- 
cellent photographs of the flash spectrum at Gallivare in 
Lapland, one of the few places where clear skies greeted the 
observers of this eclipse. The spectrograph was one of three 
prisms and slit, the photograph of the first flash was in splen- 
did definition between the wave-lengths 4154 and 4751 A. 
The dispersion of i mm equal to 3 angstroms is the largest 
ever successfully employed on eclipse spectra. 

The most interesting part of their discussion is the pho- 
tometry of the flash spectrum, the Moll registering micropho- 
tometer being used for the measurement of intensities. In 
spite of the large dispersion employed, no attempts were 
made to measure exact wave-lengths. The positions of the 
spectral lines were read from the photometer sheets and 
wave-lengths were derived from a Hartmann formula. Com- 
parisons were made for each line with Rowland and with the 
intensities estimated by Mitchell from the 1905 eclipse, an 
excellent agreement being found between the 1905 and 1927 

These spectra were very carefully calibrated by means of 
three other plates cut from the same larger photographic 
plate as the one on which the flash spectrum was photo- 
graphed, and developed together with it. On one of these 
plates was photographed the iron arc; on the second, the con- 
tinuous spectrum of a standard incandescent lamp whose in- 
tensity distribution over various wave-lengths was known; 
and on the third, the Fraunhofer spectrum of the sun. In 
front of the slit of the spectroscope they had placed a " step 
weakener." By means of this they obtained a calibration 
curve for each region of the spectrum separately; and by 
means of the spectrum of the standard lamp and the Fraun- 
hofer spectrum they obtained the relation between these in- 
dividual calibration curves. 

In a very thorough discussion they were able to find the 
number of ergs per second emitted in a unit solid angle by 
an arc of the chromosphere i' in length corresponding to each 
bright line of their flash spectrum. Combining this material 


with the theoretical intensities of lines within multiplets, 
they made a careful study of the effects of self-reversal in the 
flash spectrum. They also found the total intensities of 
entire multiplets relative to the total intensities of other mul- 
tiplets of the same element. When the theory of atomic 
spectra has been sufficiently developed to predict the relative 
intensities from one multiplet to another, such absolute in- 
tensities in the chromospheric spectrum will offer the best 
data for comparison, much better, in fact, than laboratory 

To derive from the measured transmissions of the micro- 
photometer, a knowledge of the absolute units of energy of 
the source in the chromosphere is a problem beset with many 
difficulties. By the employment of their careful methods of 
calibration, this was accomplished in three steps: (i) From 
the measured transmission curve for a short interval of 
wave-lengths to ascertain the apparent intensity. (2) By 
proper allowance for instrumental causes, such as change of 
dispersion in the spectrograph, color sensitivity of the pho- 
tographic plate, selective absorption in the apparatus, and 
other factors, to change the apparent intensities thus deter- 
mined into real intensities. (3) By proper methods of cali- 
bration and standardization it is then necessary to find the 
absolute intensities. According to Pannekoek and Minnaert, 
" It is easy to foresee that the determination of the apparent 
intensities is the most accurate, that the determination of 
the real intensities is more difficult, and that the determina- 
tion of absolute intensities may be liable to many sources of 


After making allowance for all possible factors and deriv- 
ing intensities measured in absolute units, comparisons were 
then made with Rowland's intensities and with those of 
Kayser. An interpretation of these comparisons which will 
be reasonably free from large systematic errors is encum- 
bered with many grave difficulties for the reason that the ab- 
solute scale of intensities derived from the microphotometer 
has no definite relationship with the estimates of Row- 
land and Kayser which are on arbitrary scales. On account 
of the many lines of the iron spectrum, this element gave the 


most complete information. Between the microphotometer 
readings and Kayser's estimates there was found a strong 
dependence on wave-length; a line at 4200 A having a three- 
fold greater microphotometer intensity than a line of similar 
Kayser intensity at 4700 A. On the other hand, a close 
agreement was found to exist between Rowland's solar in- 
tensities and Kayser's estimates for Fe; but a similar agree- 
ment is not found for other important elements in the sun, 
such as Cr, Mn, Ti y Ni and Co, the Rowland's intensities be- 
ing much lower than Kayser's intensities of the elements. 

In the calibration of Rowland's scale, referred to in a later 
chapter, Russell, Adams and Miss Moore find that the Row- 
land scale at 4200 A is seven per cent greater than at 4700 A, 
a result not in harmony with the findings from the 1927 
eclipse. On account of this disagreement it is evident, either 
that the 1927 spectra were not properly calibrated or else 
that the homogeneous scale derived at Mount Wilson in the 
process of the revision of Rowland's Table must be changed 
by a large factor. 

As a further result of their spectra, Pannekoek and Min- 
naert find that, " The energy emitted by the whole chromo- 
sphere in each of these wave-lengths is not proportional to 
the theoretical emission. In general, the observed intensities 
increase together with the theoretical intensities, but more 
slowly. This proves that the chromosphere may not be con- 
sidered as an optically thin layer of small effective depth, but 
that there is an appreciable amount of self-absorption in the 
layer of gas viewed tangentially." Further discussions car- 
ried out by Pannekoek 1 lead to the conclusion that the density 
of the chromospheric gases being small, the effect of colli- 
sions may be neglected with the result that absorption and 
emission of radiation is the only method of energy transfer. 
The atoms forming the chromosphere being exposed to the 
radiation from the hemisphere below, there is no diminution 
of radiation at the highest levels. This implies that the chro- 
mospheric gases produce only a small part of the total ab- 
sorption in the solar chromosphere with the consequent result 
that the Fraunhofer absorption takes place almost com- 

i B. A. N., 4, 263, 1928. 


pletely within the confines of the lower levels of the reversing 
layer. These results lead to the conclusion that the chromo- 
spheric gases are supported mainly by radiation pressure. 
For obvious reasons, these conclusions must be accepted with 

By combining the material from six successive exposures 
they were able to find the density gradient of hydrogen in 
the chromosphere up to heights of 3000 km. They found that 
its density gradient was much less than would be expected in 
an isothermal atmosphere in which the only counterbalance 
to gravitation is gas pressure. By the same method they 
found that the density of the emitting atoms of helium is 
actually greater at 700 km above the photosphere than at 
levels closer to the photosphere. Their work is the most suc- 
cessful yet accomplished in exact spectrophotometry of the 

The region investigated at the 1927 eclipse, of 600 ang- 
stroms at the red side of 4153 A, is but a small portion of the 
chromospheric spectrum. Menzel (op. cit., p. 12) with the 
Moll registering microphotometer measured the 1905 Lick 
moving-plate spectrum between 3913 and 4466 A, the region 
of best definition. Runs were made in the ordinary way at 
six different levels, assumed to be at 170,385,600, 1030, 1675, 
and 2030 kilometers, respectively, above the photosphere. 

Although the method adopted by Menzel, developed in- 
dependently, for reducing the microphotograms, was very 
similar to that of Pannekoek and Minnaert, nevertheless he 
worked under severe handicaps in that he did not even know 
the character of the photographic plate used at the 1905 
eclipse, and naturally, no standard squares were impressed 
thereon for purposes of calibration. No corrections were 
made by him for atmospheric or instrumental absorption. 

As stated in the preceding chapter, the slot of the moving 
plate permitted an integrated radiation of 300 kms in each 
spectral line, and moreover, each line under consideration 
was defined on one side by the serrated edge of the moon, 
and on the other side by emitting atoms assumed to lie above 
the six different levels under consideration. 

With so many unknown factors, and with heights subject 

From the painting by Howard Russell Butler, N. A. 


to systematic errors, Menzel is forced to evaluate his material 
from the interagreement of results from separate lines. 
"When the continuous background was subtracted, it was 
found that all of the curves for a given chromospheric height 
required multiplication by a single factor, inversely propor- 
tional to the ordinate at maximum, to bring them into sur- 
prisingly good agreement. This was probably due, in part, 
to the short spectral region studied. No correction for atmos- 
pheric or instrumental absorption has been introduced. The 
intensity of a spectral line is equal to the area under the ' true 
intensity ' curve. Since all of the curves, for a given level, 
were of similar ' shape,' the ordinate at the mid-point deter- 
mines the relative intensity of a line. 

" The observed peculiarity of line shape is just the reverse 
of what might have been expected on a priori grounds. It is 
easily shown that the widening that arose from the deviation 
of the small segment of the solar crescent from a straight line 
was negligible. Almost all other causes that act to broaden 
a spectral line pressure, Stark effect, self-reversal, tem- 
perature ( Doppler motion ) apparently should be more 
effective at the lower levels. There is, of course, the chance 
that ' seeing ' conditions during totality were not uniform, 
but there is evidence against this explanation, in the form of 
a marked exception to my previous statement that all lines, 
at a given level, conform to the same contour curve. The 
line 4026 A of helium is essentially a i high-level ' line, as is 
immediately evident from its great extension in relation to 
its intensity. Consequently, most of the intensity, recorded 
on the tracings at lower levels, is the integrated light from 
much greater heights. The lines of greatest intensity pre- 
sent a special problem. Small errors in drawing the char- 
acteristic curve of the photographic plate produce large er- 
rors in the resultant intensities. Rather than introduce such 
indeterminate intensities, I have computed the intensities of 
these lines on the assumption that their contours were the 
same as for the fainter lines. It then suffices to determine 
the ordinate at any point of the line contour. The intensity 
is measured, not at the center of the line, but in the wings 
where the photographic density is less, and where the char- 


acteristic curve is more certain. Intensities calculated in this 
manner are ' upper limits.' They are subject to correction for 
sel f -r ever sal. " 

Continuing the study by the investigation of the difficult 
problem of self -reversal, Menzel continues on page 261 as 
follows. "The density gradient of the chromosphere is 
closely related to the observed intensities of the chromo- 
spheric lines at the various levels studied by the micropho- 
tometer. The energy at any one level on the plate represents 
the integrated light from all the higher levels not yet cut off 
by the moon's limb. The presence of self -reversal compli- 
cates the problem. Theoretical estimates of the amount of 
self-reversal seem to be unreliable. We can, however, de- 
termine the amount of the self-reversal in what seems to be 
a perfectly direct way, by comparing the observed and theo- 
retical intensities through the intermediary of the estimated 
flash intensities." 

After going through an intricate theoretical discussion, 
Menzel apparently puts little weight upon his final results. 
This is shown by the fact that in the determination of the 
distribution of vapors in the chromosphere and the deriva- 
tion of density gradients, he prefers the density gradients de- 
duced from a discussion of the observed heights to those de- 
rived from his microphotometer readings. 

One of the most pressing needs concerning the atom at the 
present time (1932) is a more accurate knowledge of the 
contours of lines in the ordinary solar spectrum in order to 
ascertain the numbers of atoms effective in producing ab- 
sorption in the centers and wings of individual lines. Valu- 
able work has been done by Unsold and others, but this forms 
merely the beginning of a more comprehensive research nec- 
essary. In the Fraunhofer spectrum the atoms are absorb- 
ing energy, in the chromospheric spectrum they are emitting. 
To those who are not fully acquainted with the difficulties 
involved, it would seem an easy matter to photograph the 
chromospheric spectrum at different levels, and then to ob- 
tain a knowledge of the number of atoms involved at differ- 
ent levels. After several years of investigation on the Fraun- 
hofer spectrum, with huge equipment and unlimited re- 


sources, we do not yet feel on very firm ground. If an eclipse 
came every day, and if photographs could be obtained of the 
flash spectrum with powerful equipment, that is, with large 
dispersion and large solar image, much might be accom- 
plished. Unfortunately, eclipses are rare phenomena, the 
" flash " lasts but a few seconds, and it is impossible to use 
large dispersion with a large solar image on account of the 
paucity of light and erratic conditions of seeing. The pres- 
ent writer is always an optimist (or he would not be an eclipse 
observer) but he has a practical turn of mind. In the im- 
mediate future, he cannot foresee any brilliant discoveries 
from the measurement of eclipse spectra by the micropho- 

A knowledge of conditions in the sun's atmosphere are of 
such great importance to solar physics that attempts are 
made to observe or photograph the chromospheric spectrum 
at times when there is no total eclipse. This is possible either 
during a partial eclipse by observations at the cusps, or when 
there is no eclipse at all. Observations were secured visually 
by Fowler in London during the progress of the partial eclipse 
of the sun, April 17,1912. At its maximum phase, this eclipse 
was 0.91 total (the sun's diameter = i.o). The 6-inch equa- 
torial was employed in conjunction with the Evershed spec- 
troscope, the observing conditions being very favorable. The 
number of bright chromospheric lines that were seen vastly 
exceeded expectations. As early as thirty-five minutes be- 
fore maximum eclipse, high-level chromospheric lines were 
noted at the cusps and the number of lines increased with the 
progress of the eclipse. During the maximum phase, hun- 
dreds of Fraunhofer lines were reversed, in fact the number 
of lines seen was so great that it was impossible to record all 
of the lines. The appearance greatly resembled the flash 
spectrum that had already been observed by Fowler at more 
than one total eclipse. About seventy bright lines between 
b and D were actually identified. That so many lines of the 
chromosphere were visible near the cusps at the time of the 
eclipse may be partially explained as the result of reduced 
sky illumination due to the fact that but nine per cent of the 
sun remained uncovered. A greater advantage may have 


resulted from the smaller effect of unsteadiness or " boiling " 
at the cusp as compared with the limb under ordinary con- 

The success of these observations was so great that it 
seemed possible to Fowler that even better results could be 
obtained at a similar eclipse by the use of more powerful 
apparatus, and that the multitude of bright lines seen visu- 
ally for half an hour, while the eclipse ranged from eight- to 
nine-tenths total, might profitably be photographed by suit- 
ably arranged instruments. In fact, at this same eclipse, 
Newall at Cambridge actually secured successful photo- 
graphs. On the best photograph only about forty lines were 
recorded as bright, many of them exceedingly narrow with 
the continuous spectrum quite faint. Apparently, therefore, 
it is more difficult to photograph the bright lines than it is 
to observe them visually. In view of the success attained in 
photographing, Newall came to the conclusion ' that " ex- 
ceedingly valuable work could be carried out with an instru- 
ment of high power by an observer who, in a total eclipse, 
stationed himself to the north or south of the band of to- 
tality at such a distance that the maximum phase was about 
0.99. At such a station, detailed observations could be con- 
veniently made with a much more leisurely program and with 
far greater completeness than on the central line." 

In view of their success in 1912, the Cambridge 2 observers 
prepared to photograph the eclipse of 192 1 with the McClean 
spectroscope of the Solar Physics Observatory. This instru- 
ment uses an image of the sun 168 mm in diameter, the 
dispersion being caused by a 6-inch plane grating. Two 
photographs were secured with excellent definition but the 
total number of bright lines visible was only two, due to hy- 
drogen, not a single bright metallic line appearing on either 
of the plates. A similar disappointment resulted from the 
photographs taken with the Huggins refractor, also at Cam- 

Observations similar in kind were also made 3 visually at 

1 Monthly Notices, R. A. S., 72, 538, 1912. 

2 Monthly Notices, R. A. S., 81, 482, 1921. 

3 Monthly Notices, R. A. S., 81, 485, 1921. 


the same eclipse by Fathers Cortie and Rowland at Stony- 
hurst. A Browning spectroscope with a dispersion of eight 
prisms of 60 was employed, the region observed being in 
the green from 5167 A to 5400 A. " Every line in the field 
was reversed, the bright lines tapering to a point indicating 
decrease of pressure." 

Newall (loc. cit.) could not offer any explanation for his 
failure to photograph the bright lines at the 1921 partial 
eclipse. The experience of the writer, both at total eclipses 
and at Mt. Wilson (see below) causes him to suspect that the 
conditions of seeing at Cambridge were inferior to what they 
were at the eclipse of 1912 in spite of the fact that at the 
time the observers regarded the conditions as about equal. 

A quarter of a century ago, 1 Hale and Adams secured pho- 
tographs of the flash spectrum without an eclipse by using the 
6o-foot tower telescope and powerful spectograph attached. 
The method employed was to allow light from the sun's limb 
to fall upon a diagonal prism so placed as to reflect the light 
horizontally to a second prism immediately above the slit. 
The first prism was mounted upon a slide with a screw ad- 
justment allowing motion toward or away from the second 

After the sun's image has been brought to the slit, the 
observer selects a bright line of the chromospheric spectrum 
and brings it into the field of view of an eye-piece, mounted 
in an opening near the end of the plate-holder. During the 
exposure, this line is maintained at maximum brightness by 
guiding with the screw controlling the position of the first 
diagonal prism, thus moving the sun's image slightly on the 
slit. The second order spectrum was employed giving a dis- 
persion of i mm = 0.9 angstroms; and exposures of four 
minutes were required in the yellow part of the spectrum, 
and double this amount in the red. 

Two reports on the wave-lengths, etc., obtained from the 
spectra have been published from Mt. Wilson. The first one 
by Hale and Adams appeared in Communications No. 41, 
and the second by Adams and Burwell in Communications 
No. 95, the former coming before and the latter after the 

1 Astrophysical Journal, 30, 222, 1909; Mt. Wilson Contr., No. 41. 


publication by Mitchell of his flash spectrum results obtained 
at the eclipse of 1905. For obvious reasons, the second com- 
munication from Mt. Wilson is more complete than the first. 
Unquestionably the photographs at Mt. Wilson were secured 
at a very low level, but whether at a lower level than the 
1905 eclipse spectra it is rather difficult to decide. In com- 
paring the results obtained within and without an eclipse, 
there are very great advantages in favor of the latter method, 
the most important being the possibility of securing a much 
higher dispersion by the method without an eclipse than can 
be secured at the time of an eclipse. The dispersion of the 
1905 eclipse spectra was i mm =10.9 angstroms, or one- 
twelfth of that in the discussion in the Mt. Wilson Com- 
munications. For reasons stated in the foregoing, it is im- 
possible to use large dispersion at an eclipse. Because the 
exposures without an eclipse can be increased at will, the dis- 
persion can also be increased, and it may be possible for the 
Mt. Wilson observers to employ the i5o-foot tower tele- 
scope with the 7 5 -foot spectograph in the second order and 
thus secure a dispersion thirty times that of the 1905 eclipse 
spectra. At an eclipse only two exposures of the flash spec- 
trum are possible, and unfortunately the best focus is se- 
cured more or less only as the result of a happy accident. 
Outside of an eclipse there are no such limitations, observa- 
tions may be repeated until moments of good seeing are se- 
cured and until photographs are obtained of the best defini- 
tion exhibiting the results of layers sufficiently close to the 

The comparison of the results obtained with and without 
an eclipse may be briefly summarized here. First, let us 
take up the number of the lines photographed by the two 
methods in a given region of the spectrum. Compared with 
spectra taken outside of an eclipse, those obtained at an 
eclipse have the two-fold disadvantage of the much smaller 
dispersion and of the superposition of the eclipse lines on 
a very strong continuous spectrum. In the process of meas- 
uring the eclipse spectra, it was difficult to differentiate the 
weakest lines from the continuous spectrum. Consequently 
many eclipse lines were actually measured in the flash spec- 


trum but were not published at the time (see Astro physical 
Journal, 38, 481, 1913), for it seemed unwise to add any lines 
to the publication unless those lines could be definitely identi- 
fied with lines appearing in Rowland's Tables. For these 
reasons, therefore, it must be concluded that the number of 
the weakest lines secured at Mt. Wilson without an eclipse 
is probably not greater than the number found on Mitchell's 
eclipse spectra. Since the intensities of the lines depend on 
the levels at which the lines originate, it seems highly prob- 
able, therefore, that the levels photographed within and with- 
out an eclipse are not very different. For the stronger lines 
originating at higher levels, the intensities inside and outside 
of the eclipse greatly differ, such variations however being 
readily explained on account of these very differences in level. 

Regarding the accuracy of the wave-lengths, the following 
may be said. On account of the spreading of the photographic 
images of the strong lines of the eclipse spectra it is im- 
possible to secure accurate wave-lengths from such lines. 
Omitting these lines, the eclipse wave-lengths differ from 
Rowland on the average by 0.020 angstroms. With the 
twelve-fold greater dispersion at Mt. Wilson the wave- 
lengths have an accuracy of 0.012 angstroms, 1 being a pre- 
cision about twice that obtained from eclipse spectra. 

With a slit tangent to the sun's limb it is evident that the 
length of the lines in the spectrum taken without an eclipse 
should furnish information regarding the levels at which 
such lines originate. Such measures have not as yet been 
carried out at Mt. Wilson. The need of more reliable deter- 
minations of the levels at which the spectrum lines of dif- 
ferent intensities originate and the great importance of such 
knowledge in present-day problems of solar physics cannot 
be overemphasized. 

In summing up the problem regarding the photography of 
the flash spectrum as it appeared to him in 1924 while in resi- 
dence for a brief time at Mt. Wilson Observatory, the writer 
voiced his opinions in the following paragraph which is found 
in the second edition. 

The recommendation is therefore made to the investiga- 

1 Mt. Wilson Contributions, No. 95. 


tors of solar rotation, and is hereby urged upon them for 
their consideration, that they curtail their work on solar 
rotation, and instead devote their energies for a short while 
to the flash spectrum without an eclipse. With the equip- 
ment already in hand, and under good conditions of seeing 
and with adequate facilities for proper guiding, good photo- 
graphs of the flash spectrum should be possible at very low 
levels. The accurate measurement of wave-lengths, and par- 
ticularly the determinations of the levels at which different 
lines of the spectrum originate, will add greatly to our knowl- 
edge concerning these lines and will supply information so 
sadly needed in deciphering the laws underlying the produc- 
tion of spectral lines. The Bohr atom and Saha's theory of 
ionization have made possible the identification of new 
spectral lines in the sun and in the laboratory, and undoubt- 
edly we are on the eve of a very great advance in knowl- 
edge regarding the laws underlying the production of lines 
in the spectra of various sources, provided wave-lengths 
of great accuracy and free from systematic errors can be 

The eight hectic years since the second edition of this book 
have revolutionized our conceptions of solar research. Re- 
garding the problem of the flash spectrum, Director Adams 
gave the writer, while he was in residence at Mt. Wilson, 
every opportunity for observing with the i5o-foot tower tele- 
scope. The writer then realized the important role played 
by the qualities of seeing and steadiness of image of the sun 
on the slit of the spectrograph. If the slit is set tangent to 
the limb of the sun and if the bright-lined spectrum is ob- 
served visually by the methods described above, then it is 
immediately noticeable that if the seeing is ragged or fair, 
there will be few bright lines that can be seen, such for in- 
stance as the stronger lines of the solar spectrum like the 
6-group in the green. If the seeing improves and the image 
of the sun becomes more and more steady, then more and 
more lines are seen reversed into bright lines. At Mt. Wil- 
son, the best seeing ordinarily comes shortly after sunrise. 
After the sun's rays heat up the slopes of the mountain and 
the rising of the heated air disturbs the atmosphere, the 


steadiness of the image of the sun on the slit of the spectro- 
graph becomes less and less. In the short time at his disposal 
for making observations, the writer came to the conclusion 
that for photographing the flash spectrum without an eclipse 
the best possible conditions are absolutely necessary, or in 
other words, with seeing and steadiness of quality 10 where 
10 represents the maximum of perfection. If therefore one 
wishes to obtain such photographs without an eclipse, he 
must be continually on the alert to catch the few precious 
moments of perfect seeing when they come; and then if not 
too impatient, after taking hundreds or possibly thousands 
of photographs for this purpose, a photograph may finally be 
secured which will be of quality equal to, or even better than, 
that of Adams obtained in the early years at Mt. Wilson. 
These observations only serve to emphasize the importance 
of excellent seeing, no matter what method is employed for 
photographing the flash spectrum. 

Anticipating somewhat the theoretical considerations to be 
developed in the three following chapters, it is possible to 
make a summary of results achieved from a study of the 
flash spectrum, and also to make recommendations for work 
at future eclipses. The following general conclusions may 
be drawn: 

1. On account of the different conditions under which the 
chromospheric and Fraunhofer spectra originate, it is prob- 
able that wave-lengths from the two spectra differ systemati- 
cally. However, these differences do not amount to as much 
as 0.02 A. 

2 . Every strong line in the Fraunhofer spectrum is found 
in the flash spectrum, and every strong line in the latter 
(with the exception of H and He lines) is matched by a line 
in the former. 

3. The flash spectrum may therefore be regarded as a 
reversal of the Fraunhofer spectrum. 

4. The " flash " is not an instantaneous appearance. At 
the beginning of totality the chromospheric lines of greatest 
elevation appear first, and at the end of totality remain the 

5. The " reversing layer " which contains the majority of 


the low-level lines of the chromosphere is about 600 km in 

6. The " reversing layer " has no existence separate from 
the chromosphere. 

7. It is the densest part of the chromosphere lying closest 
to the photosphere, and it is the cause of the greatest por- 
tion of the absorption producing the Fraunhofer lines. 

8. The " Evershed-effect " measured in sun-spots, and 
photographs of flocculi which exhibit vastly different aspects 
when photographed at various elevations above the photo- 
sphere prove that the shadings of such strong lines as H and 
K are caused by absorption at different levels and pressures 
above the photosphere. 

9. The chromospheric spectrum differs greatly from the 
ordinary solar spectrum in the intensity of the lines. 

10. The Fraunhofer spectrum is essentially an arc spec- 
trum. The chromospheric spectrum more closely resembles 
the spark spectrum, and its spectrum corresponds to an 
" earlier " type than that of the sun. 

1 1 . Especially prominent in the chromosphere are the en- 
hanced lines. 

1 2 . The enhanced lines ascend to greater elevations above 
the photosphere than do the ordinary lines. 

13. The increased elevations cause greatly diminished 

14. As Saha has shown, the reduced pressures permit the 
ionization of the atom. As a result, the lines of the ionized 
atoms are specially prominent in the flash spectrum. The 
enhanced lines are produced by the ionized atoms. 

15. The depth of the chromosphere is not constant. 
Recommendations for future work on the flash spectrum 

may be briefly summarized. The most important contribu- 
tion from the chromosphere will undoubtedly come from the 
investigations of heights and intensities of the spectral lines 
in order to gain further information regarding atomic struc- 
ture. To be of the greatest value in furthering scientific 
research, photographs of the flash spectrum should be se- 
cured with large dispersion, they should extend as far to 
the violet and as far to the red as possible, the definition 


should be of the very best and the exposures should be 
timed so as to photograph the very lowest possible levels. 
The occupation of a station near the edge of the moon's 
shadow path would permit relatively long exposures on the 
low-lying layers closest to the sun's pole. Such spectra would 
afford interesting comparisons with those taken near the cen- 
tral line of totality which give spectra near the sun's equator. 
Another important investigation for the future will be to 
make comparisons between flash spectra taken at different 
phases of the sun-spot period. If we are to judge by the 
changes which take place in the spectra of the corona, we 
should expect that the flash would be richest in lines at times 
of maximum sun-spots. The eruptions taking place near 
spot zones seem to elevate the low-lying metallic vapors. If 
this be true, it would be natural to expect that if it were pos- 
sible to photograph at a constant level above the photo- 
sphere in securing the flash spectrum at an eclipse, the lines 
should appear to be of greater strength at sun-spot maximum 
than at minimum. 

The type of spectrograph is all important. Eclipses, how- 
ever, are rare phenomena. After the eclipse of 1932, which 
will be assiduously observed, the total eclipses of the next 
two decades require long voyages and expensive expeditions. 
Few institutions, except those like the Lick, U. S. Naval and 
Mt. Wilson Observatories in the United States, and the 
British Joint Permanent Eclipse Committee, can afford the 
luxury of assembling spectrographs for infrequent use on 
chromospheric work. Hence most astronomers will be forced 
to use the equipment they already have, others, less fortunate, 
must borrow their apparatus. 

As it is impossible to employ a dispersion sufficiently high 
to permit the detection of systematic differences in wave- 
lengths between the solar and chromospheric spectra, we 
come to the conclusion that eclipse wave-lengths are of sec- 
ondary importance, serving as they do merely as a means of 
accurately identifying the origins of the spectral lines. Al- 
though it is far easier to get good definition by the use of a 
slit spectrograph, each observer must come to the decision 
whether to try the easier method and thereby sacrifice in- 


formation regarding levels, or to try the more difficult plan 
of photographing without a slit in order to attempt to 
gain knowledge regarding heights which are now about the 
most important contribution that can come from eclipse 

In work on eclipses, as in all other branches of scientific 
research, continued advances must be made, if the time and 
energy and money are to be profitably spent. Just because 
eclipses are rare phenomena is no reason in itself why an 
astronomer without eclipse experience can expect to make a 
scientific contribution of value unless he knows the subject 
thoroughly and has a problem of importance to attack. 

Prisms have advantages over gratings in their greater 
light-gathering power. For work at the blue end of the spec- 
trum, prisms may be the best means of securing the high 
dispersion now necessary. On account of erratic conditions 
of seeing at the time of an eclipse, no one should attempt to 
increase the dispersion by increasing the focal length of the 
camera to 20 feet or more. No photographs of the flash spec- 
trum of first quality have yet been taken by this type of 
prismatic camera. In the future, its use at eclipses should be 

It is possible to use higher dispersion on the chromosphere 
than has been employed up to the present. We badly need 
photographs at the extreme violet end of the spectrum with 
large dispersion in the region which cannot be reached by 
glass prisms, and hence quartz must be used. 

Except for the light-gathering power, gratings have dis- 
tinct advantages over prisms. At the red end the prismatic 
spectrum is crowded together. To get high dispersion, many 
prisms must be used. The grating gives a " normal " spec- 
trum and a greater region of wave-lengths in good definition. 
If gratings are used, they should be concave and not flat grat- 
ings. Similarly, larger dispersion than that employed to 
date may be used profitably with concave gratings. To in- 
crease the dispersion one cannot expect to have more lines 
per inch than the number of 15,000 per inch which have been 
used. The only other means is to work in the second order 
(if it is possible to rule gratings specially for this purpose) , or 


to increase the radius of the grating beyond that of 10 feet, 
the maximum so far successfully used at eclipses. As Ander- 
son L has pointed out, if the seeing is not of the best at the 
time of the eclipse, the area of the chromospheric arcs on the 
photographic plate will be increased with increase of focal 
length, with the consequent diminution in the intensity of 
illumination. As a result, the faint chromospheric arcs will 
be below the threshold value of the plate and, in consequence, 
these lines will be conspicuous by their absence. To over- 
come this difficulty, Anderson put into practice for the 1926 
eclipse the use of a 21 -foot concave grating of 15,000 lines 
per inch, but the sizes of the chromospheric arcs were di- 
minished by the employment of mirrors. This method has 
been tried out at the eclipses of 1926, 1927 and April, 1930, 
but clouds interfered in each case. It will be tried again in 

The author does not believe that anyone will be brave 
enough, or perhaps foolhardy enough, deliberately to place 
himself outside the path of totality in order to photograph 
the spectrum of the chromosphere. At the eclipse of 1900, 
Evershed found himself outside the eclipse track as the re- 
sult of an accident. To have the longer time of observation 
by observing at the cusps, it will be necessary to await the 
time when a total or annular or large partial eclipse passes 
near an observatory where there is a powerful equipment 
with large dispersion. Observations of value will be pos- 
sible only under the condition that the eclipse is more than 
ninety-five per cent total. The present generation of active 
astronomers will not see such conditions come to pass within 
their lifetimes. At the 1932 eclipse, the physical laboratory 
of McGill University will be inside the southern edge of the 
eclipse path. 

If progress is made in the near future on the spectrum of 
the chromosphere with a much greater dispersion than is 
possible at eclipses, the only method left is that without an 
eclipse. Unfortunately, this work demands very high quali- 
ties of seeing that practically never occur. Hence, taken all 
in all, it appears probable that we shall have to rely on total 

1 Publications A. S. P., 38, 239, 1926, 


eclipses for adding to the information already available on 
the chromosphere. 

The splendid work of Adams at the Mt. Wilson Observa- 
tory in photographing the spectra of the brightest stars of 
various types with very high dispersion will supplement the 
work on the chromosphere by giving information on series 
relationships and multiplet groups in spectra. The work 
so beautifully inaugurated with the powerful spectrograph 
attached to the loo-inch mirror may be carried out with 
greater facility when the gigantic 2oo-inch telescope is put 
into operation. 

At the present time, astronomers are devoting much atten- 
tion to the measures of intensities by means of the micro- 
photometer. In the foregoing, the author has tried to point 
out some of the difficulties to be overcome when applying pho- 
tometric measures to eclipse spectra. Unless the measures 
can be reduced to absolute units, fairly free from systematic 
errors, the measured intensities will have a reliability little 
higher than estimates by a skilled observer. In all of the 
discussions so far published, the intensities of the chromo- 
spheric lines have been compared with Rowland's values and 
with those of the laboratory, both of which are on estimated 
scales. Hence when reducing eclipse spectra, each observer 
must decide for himself whether it is worth while both to 
estimate and measure chromospheric intensities, when per- 
force all comparisons must be made with arbitrary estimates. 
However, it is very desirable that each eclipse spectroscopist 
should place standard squares on his photographic plates so 
that later these may be available for photometric purposes if 

Fixed or moving plates may be used for the flash spectrum 
with either prisms or grating. Those who disagree with the 
author in his opinion that the fixed plate gives the more reli- 
able information have the opportunity of using the moving 
plate method. At the 1932 eclipse, both methods will have 
a thorough trial for there will be many American and Euro- 
pean parties in the field with well-equipped expeditions. 



THE discussion of the results of the flash spectrum 
obtained in 1905 has shown the great importance of 
enhanced lines, or those due to the ionized atom, 
and the explanation offered by the writer was that the cause 
of the enhancement was due to the great heights attained by 
the vapors producing the enchanced lines and, consequently, 
to the reduced pressures at which these lines were found. 
It was also pointed out that at the time of an eclipse the 
light from the sun is capable of reaching us only in a direc- 
tion tangential to the sun's surface. As a result, a beam of 
light from the bottom of a layer only 500 km in thickness 
would encounter no less than 20,000 km of emitting atoms 
in the tangential line of sight before getting out of the 
shallow layer. Any theory that can explain the method 
whereby atoms are ionized will be of the very greatest im- 
portance in all problems of modern astronomy. Such a 
theory has been developed by Dr. Megh Nad Saha. The 
sun is the nearest of the fixed stars, and on account of its 
proximity its structure may be examined in detail. In the 
sun and stars are found high temperatures, very minute 
pressures and electromagnetic conditions that together cause 
ionization to take place with great facility. The celestial 
laboratories thus opened to the astronomer have given him 
the opportunity of supplementing the work of the physicist 
and chemist in terrestrial laboratories in the combined attack 
on atomic structure. 

Accepting the correctness of the Bohr-Sommerfeld theory 
of atomic radiation, and assuming that the general laws of 
thermodynamics apply equally well to electrons and to mole- 
cules of gases, Saha has been able to calculate the degree 
of ionization that takes place in gases under different con- 



ditions of temperature and pressure, and has derived formu- 
las which can readily be applied to conditions existing in the 
atmosphere of the sun and of the stars. This theory ex- 
plains both qualitatively and quantitatively many of the 
features observed in the spectrum of the sun and of the stars, 
and it likewise finds a ready application in laboratory spectra 
under conditions when enhanced lines appear. 

In addition to the original papers by Saha, 1 valuable con- 
tributions have been made by Milne, 2 by Russell," and also 
by many others. 4 Assuming that the decomposition of a 
molecule or an atom into one or more electrons and a posi- 
tively charged ion is essentially of the same nature as an 
ordinary chemical reaction, Saha derives a simple equation 
to express the self-ionization of a gas at high temperatures. 
The equation derived is: 

P x 2 

I X 2 

x/(i x} is the ratio of the percentage of atoms ionized 
to those left neutral, and this ratio multiplied by the partial 
pressure of the free electrons (P x/(i + x) ) is equal to K y 
which is a function only of the absolute temperature of the 
gas and the ionization potential. This latter is a measure 
of the work done to ionize a single molecule, or to drive an 
electron from its neutral ring to infinity, and it is expressed 
as the number of volts through which the electron must 
fall to acquire this energy. Since the ionization potential is a 
constant, the quantity K, in the formula above, depends only 
on the absolute temperature. Hence for a given pressure, 
the smaller the ionization potential P, the more nearly x 
approaches unity, or in other words the more nearly com- 
plete is the ionization. For all gases where the ionization 
potential is known, Saha is enabled to calculate the per- 
centage of ionization found under different conditions of tem- 
perature and pressure. The higher the value of the ionization 

1 Philosophical Magazine, 40, 472 and 809, 1920; Ibid., 41, 267, 1921; and 
Proceedings of the Royal Society, A, pp, 135, 1921. 

2 Monthly Notices, R. A. S. Many papers beginning 83, 1923. 

3 Astrophysical Journal, 55, 119 and 354, 1921. 

4 Eddington, Internal Constitution of the Stars, 1926. 

Slit set at 







AUGUST 25, 1904 

The slit was set at different wave-lengths corresponding to different levels above the 
sun's surface. Order: from lowest upwards. 

The 40-foot camera is placed horizontally. 

The 40-foot camera is pointed directly at the Sun, 



potential, the higher must be the temperature to sustain a 
given degree of ionization. This is readily seen in the case of 
helium which possesses the highest known ionization poten- 
tial, of 24 volts ; for the series due to enhanced helium is found 
only in stars of the highest temperature. 

Saha calculates tables giving the percentage of ionization 
in atmospheres at various temperatures and pressures ( meas- 
ured in atmospheres). The following tables are copied from 
his publication. 


Pressure fatmos ) 





10 - 































4 6 


g8 5 






g8 5 




1 2,000 


9<> 5 













Pressure (atmos ) 



IO" 1 

















7 *> 


2 2 



io r ooo 


9<S 5 

I 2,000 


97 5 







10 4 






















Saha applies his theory to explain the differences between 
the spectrum of the sun and the chromosphere, and Russell 
broadens the scope of the theory by applying it to show the 
meaning of the differences in intensities of lines in the sun- 
spot and in the solar spectrum. The results furnish a com- 
plete triumph for the Saha theory. 


Fortunately for the theory, the flash spectrum gives the 
height in kilometers (or miles), that the various vapors 
producing different spectral lines extend above the level of 
the sun's photosphere. Within the past few years there has 
been a great revision of ideas regarding the pressures found 
in the reversing layer. Where formerly it was regarded that 
these pressures amounted to several times that found at 
sea level on earth, it is now known with certainty that the 
chromospheric pressures are even less than a thousandth of 
an atmosphere. According to many researches in recent 
years by a number of competent investigators, 1 it is known 
that the solar temperature is not far from 6000 C. abso- 
lute. Schwarzschild ~ has shown that if the variation in 
temperature in the upper atmosphere is caused only by radia- 
tion, then the temperature should not fall below 6ooo/2*, 
or 5000, approximately. Hence in the solar atmosphere, 
where the temperatures vary between 5000 and 6000, it is 
easy to see at a glance from Saha's tables the percentage of 
ionization. For instance, in the chromosphere at elevations 
where the pressure amounts to one ten-thousandth of an at- 
mosphere, ionization for calcium is ninety per cent complete. 

Saha's theory is in complete harmony with the conclusions 
derived from the discussion of the flash spectrum and fur- 
nishes an adequate explanation. The case of calcium is 
most interesting. The lines H and K are enhanced lines and 
are caused by the ionized atom, while the g line at 4227 A 
takes its origin from the neutral atom. In the neighborhood 
of the reversing layer, both normal and ionized atoms will 
be plentiful and the presence of g and the H and K lines are 
fully explained. At great heights above the reversing layer, 
however, the pressure will be very small and, as a result, 
ionization will be nearly complete. Under these conditions 
the neutral atom cannot exist, the ionized atom exhibiting 
the enhanced lines alone being found. In the flash spectrum, 
measures indicate that the H and K lines extend upwards 
to heights of 14,000 kms, but the g line only to 5000 kms. 
The presence of H and K above the 5000 km level shows 

1 Russell, Dugan and Stewart, Astronomy, 535, 1927. 
- Gott. Nachrichten, 41, 1906. 


that calcium actually exists above this level, and we must 
therefore interpret the failure of the atom to emit the g 
line above the 5000 km level to be due to the fact that prac- 
tically all of the atoms are ionized and there are few normal 
atoms left to produce the 4227 line. For strontium and 
barium, which also exhibit enhanced lines, their ionization 
potentials (5.7 and 5.1 volts) are lower than that of calcium 
(6.1 volts), and consequently complete ionization in the 
chromosphere is found at higher pressures or, in other words, 
at lower elevations above the photosphere. The strongest 
line of neutral Sr is 4607 A, which reaches an elevation of 
only 400 km, while the ionized atom Sr+ shows the two 
lines at 4215 A and 4077 A which are found at elevations 
much greater than that of the neutral atom, in fact, at 6000 
kms, but this level is much less than the 14,000 km height 
attained by the H and K lines of C0+- 

In addition to Ca and Sr, Ba also belongs to the alkali 
earths of Group II of the periodic table of the chemical 
elements. In the following table, there are given the details 
concerning the strongest single line found in the neutral series 
of each of the three elements, and also the strongest doublet 
belonging to the ionized atom. In the second column is given 
the wave-length and in the last the heights in kilometers 
measured in the flash spectrum. In the four other columns 
are given the intensities in sun, chromosphere, arc and spark, 
respectively. The intensities for the laboratory spectra are 
taken from Exner and Haschek. In spite of its high atomic 
weight ( 137.4), Ba is a conspicuous element, both in the sun 
and in the chromosphere, on account of the enhanced lines. 
No neutral lines of Ba are found either in the sun or in the 

This small table contains some of the conclusions derived 
from a study of the flash spectrum which, on account of 
their importance, will bear being repeated. The strongest 
line in the Fraunhofer spectrum is K (Rowland intensity 
1000) and this line is the strongest in the flash spectrum 
(intensity 200 on a different scale). The enhanced lines 
which are produced by the ionized atom (designated by +) 
have an intensity in the chromosphere greater than in the 



sun, and these enhanced lines extend to greater elevations 
above the photosphere than do the ordinary or unenhanced 

In the flash spectrum the elements arranged according to 
the intensity of the strongest lines give the following se- 
quence: Ca, H, He, Ti y Mg, Sr, Sc, Cr, Ba, Fe, A I and Y. 
This order differs materially from a similar table for the 








in km 


4^27 (g) 







303? (K> 







3Q6g (H) 







































Ba + 







ordinary solar spectrum. In lists arranged according to the 
number of lines, Fe stands first in each case. Ni and Co 
are second and sixth in the solar spectrum, but become 
sixth and eighth, respectively, in the flash spectrum. V is 
the eighth element in the Fraunhofer spectrum but fourth 
in the chromosphere. Cb, Mo and Pd appear in the first 
list but not at all in the second, while in the chromosphere 
is found He which is absent in the ordinary spectrum. The 
rare earths are represented relatively by more lines in the 
flash than in the ordinary spectrum. 

Interesting results are obtained when the total number 
of lines identified in the flash spectrum are arranged accord- 
ing to the periodic tables of the elements. This is given in 
the table on page 306. 

For each element in the table there is given: in the first 
line the atomic number and the chemical symbol, in the 
second line is found the atomic weight, and in the third line 
in italics the total number of lines in the chromosphere. 


The mark (?) is found with the elements Cb, Mo, Ag and 
Cd to signify that there are no strong lines in the chromo- 
sphere resulting from these elements, but they may possibly 
be represented by weak lines in combination with stronger 
lines of other elements. Eight of the rare earths are repre- 
sented in the flash spectrum by a total of 1 78 lines, Ce, La 
and Nd being responsible for 57, 47 and 31 lines, respectively. 
All of the rare-earth lines without exception are enhanced 
lines. The element Ca occupies an interesting place in the 
table. To the right of Ca in the table are found Sc, Ti, V, Cr, 
Mn, Fe, Co and Ni, all represented by numerous strong lines 
both in chromosphere and in Fraunhofer spectrum. Verti- 
cally above Ca is Mg, and below it are Sr and Ba, all impor- 
tant elements in solar investigations. In column I, of the 
alkali metals, the only element represented in the chromo- 
sphere is Na, where it is found with less prominence than in 
the Fraunhofer spectrum. He is the only element in column 
O, the inert gases, that manifests itself in the flash spectrum. 

As mentioned above, the chromospheric and solar spectra 
agree exactly as to wave-lengths, but differ very greatly in 
the relative intensities of the lines. These differences of 
intensity are accentuated in the case of the " enhanced " 
lines, or those which are more intense in the spectrum of 
the spark than in the arc. 

A comparison of the intensities of the lines in the Fraun- 
hofer, chromospheric, arc and spark spectra forces one to 
the conclusion that while the Fraunhofer spectrum corre- 
sponds to the arc spectrum, the spectrum of the chromo- 
sphere more closely resembles that of the spark spectrum. 
The sun thus exhibits three distinct spectra under different 
conditions: the chromospheric spectrum, the Fraunhofer 
spectrum and the sun-spot spectrum. These three closely 
resemble the spectra of the stars y Cygni, Capella and Arc- 
turus, respectively; y Cygni representing an " earlier " and 
Arcturus a " later " type of spectrum than that of the sun. 
Hence, the Harvard classification of the three spectra are: 
chromosphere = Fo, sun = Go and sun-spot spectrum = Ko. 

A comparison of the spectra of the sun and the chromo- 
sphere reveals the following interesting facts: first, the 





</5 C 

H ** 

g -s 

1 -S 

r^ rt 

f T ] |Jj 

a c: 


CO ^ 

* *-" 

r , ^j 

PH 3 



10 O 

<N in t-s 


cc "2 

L. O 



-t a 



s: x 

^ -t 
c ^ 



9 ^ 


I- O 

-t >-l 

'0 VJ 


values of Rowland, giving the intensities of the lines of the 
ordinary solar spectrum, are comparable with arc intensities, 
while those of the chromosphere approximate more closely 
to the intensities of the spark; and second, the enhanced 
lines of any element are in every case stronger in the chromo- 
sphere than the lines of the neutral atom, and the heights 
attained are much greater. The results are a complete veri- 
fication of Saha's theory that enhanced lines are caused by 
ionization which becomes more and more complete in the 
sun's upper atmosphere where the great altitudes make the 
pressures very small. 

Not only is Saha's theory able to explain the facts regard- 
ing the enhanced lines of the ionized atom, but it makes 
clearer the details concerning the lines of the neutral, or 
un-ionized atom. Take, for example, the D-lines of sodium, 
so well known in the Fraunhofer spectrum. At pressures 
below one-thousandth of an atmosphere, Na with an ioniza- 
tion potential of 5.1 volts, is completely ionized. The D- 
lines belong to the principal series of the normal atom, and 
accordingly, they have no connection with, and are not pro- 
duced by, the ionized atom. The normal atoms forming the 
D-lines therefore cannot exist when the pressure in the 
chromosphere is reduced to the thousandth of an atmos- 
phere. It is quite in keeping with theory to find the flash 
spectrum photographs furnishing the information that the 
D-lines reach the comparatively small heights of only 
1500 kms above the photosphere (H and K are found at 
14,000 kms). The contrast in behavior in passing from 
the Fraunhofer to the flash spectrum for the D-lines of 
sodium on the one hand, and D t of helium on the other, 
is very marked. The sodium lines are weakened in the 
flash while the helium line is enormously strengthened, 
being entirely lacking in the ordinary solar spectrum. Fur- 
thermore, in view of the great prominence of the D-lines in 
the solar spectrum, it has always been a matter of the great- 
est surprise that the element potassium, so similar in its prop- 
erties to sodium, is not found represented by strong lines in 
the sun. The explanation is a very simple one. The lines 
of the neutral atom of potassium, corresponding in its series 


to the D-lines of sodium, are found in the deep red part of 
the spectrum at wave-lengths 7664 A and 7699 A, and con- 
sequently they are not in the visible spectrum. Like the 
D-lines, both lines of this pair are strengthened in sun-spots. 
The only lines due to potassium found in the visible solar 
spectrum are very weak lines at 4044 A and 4047 A, of 
Rowland intensities o and oo, respectively. Russell finds 
both these lines strengthened in sun-spots. No enhanced 
lines are known for Na or for K, and consequently neither 
element is conspicuous in the flash spectrum. 

The temperature of the photosphere is approximately 
6000, while that of sun-spots is lower and probably some- 
where near 4700.' The pressures found in sun-spots can 
differ but little from those in the lowest depths of the re- 
versing layer. On account of the lower temperatures in the 
spots, however, ionization is less complete according to 
Saha's theory. As a result, the lines of the neutral atom, 
the so-called " low temperature " lines, are strengthened in 
sun-spots, while on the other hand, and also as a direct con- 
sequence of Saha's theory, the enhanced, or u high-tempera- 
ture " lines are weakened in the spectrum of sun-spots. 
Since the variations in pressure in the neighborhood of the 
sun are much greater than the variations in temperature, 
it would have been more fortunate if the enhanced lines 
had been referred to as "low-pressure" rather than as 
" high-temperature " lines. 

In the light of Saha's theory, Russell has investigated the 
sun-spot lines. His conclusions for the alkali metals (loc. 
cit., p. 129) are here briefly given. Sodium is represented 
in the sun by the principal, diffuse and sharp lines of the 
neutral series, and all of its lines are much strengthened in 
the spot spectrum. Potassium is represented by the principal 
series only and its lines are also strengthened in sun-spots. 
Lithium is found in the spot spectrum only at wave-length 
6707 A. Rubidium, hitherto unknown in the sun, was dis- 
covered to be present by both members of the strongest 
pair of the principal series, the wave-lengths being 7800 A 
and 7947 A. 

1 Miss Moore, Astro physical Journal, 75, 222, 1932. 


We are now in a position to explain some of the peculiari- 
ties regarding the appearance of lines in the spectrum of the 
sun and chromosphere and the heights found in the flash 
spectrum. The peculiarities noted (Chapter XVI) are as 
follows: The H and K lines of calcium of atomic weight 40 
are stronger in sun and chromosphere and reach greater 
heights than hydrogen, the lightest gas in the sun. In 
the chromosphere the whole Balmer series for hydrogen is 
found, while only the first four members are seen in the 
Fraunhofer spectrum. No helium lines are found in the 
ordinary solar spectrum, but they are of great strength in 
the chromosphere. The elements, other than H and He, ar- 
ranged according to the periodic table (see page 306) have 
remarkable progressions in the number and intensities of 
the lines involved. Group II, the alkali earths, represent the 
strong lines in the chromosphere, the strongest lines of 
all belonging to Ca. Group I, the alkali metals, have few 
strong lines in sun or chromosphere other than the D-lines 
of Na. None of the lines of Group O originating from the 
inert gases Nc, A, Kr and Xe are found in sun or chromo- 
sphere. In Group III, strong lines are found for Al y Sc, Y, 
and the rare earths, but the strength of lines is not as great 
as reached by the corresponding elements in Group II. In 
Group IV intensities are still less. The only element in 
Group V, found with certainty in the chromosphere, is va- 
nadium, and in Groups VI and VII, O and Cr, and Mn, re- 
spectively, and in Group VIII, the three metals Fe y Co, 
and Ni. 

It is easy to see why the metals of Group I, the alkalis, 
are represented by such feeble lines in the chromospheric 
spectrum. For reasons already stated, the enhanced spectra 
of the alkali metals resemble the spectra of the neutral atoms 
in the preceding group in the periodic table, the inert gases; 
and consequently, such spectra are very difficult to produce 
on account of the outer electrons forming part of a very 
stable ring or shell. As a matter of fact, no enhanced lines 
are found for any of the alkali metals in the visible portion 
of the spectrum. It is apparent, therefore, why the alkali 
metals cannot be prominently represented in the chromo- 


sphere since the flash spectrum is essentially an enhanced 

Quite different is the situation regarding the elements of 
Group II, the alkali earths, which are specially important 
in the chromospheric spectrum, for the reason that the 
strongest lines of their spectra are enhanced lines, and the 
principal members ( is 2p] of the series lie in the familiar 
portion of the spectrum. This is true for the elements with 
the exception of Mg, the strongest lines of which are found 
in the extreme ultra-violet at 2795 A and 2802 A, in a region 
in fact where no light can reach the earth's surface from 
the sun on account of the absorption of light in the earth's 

The great strength of the H and K lines of calcium both 
in the sun and in the chromosphere and the great heights to 
which these lines extend in the flash spectrum are now 
completely explained as the result of Saha's theory. In 
spite of the great difference in the atomic weights of the 
two gases, calcium and hydrogen, the atomic weight of the 
former being forty times the latter, the spectrum lines of 
calcium are seen to reach greater heights than are attained 
by hydrogen. The reasons for this curious circumstance 
are very simple. H and K are lines due to the ionized atom, 
and in virtue of the great elevations the ionization is greatly 
increased. The lines H and K are the chief lines belonging 
to the principal series, and in fact are the only lines of this 
series in the chromosphere. Only five others lines of Ca+ 
are found in the chromosphere, they are all of subordinate 
series and they have much smaller intensities than H and 
K. The hydrogen lines in the visible spectrum, on the con- 
trary, belong to a subordinate series and not to the princi- 
pal one. 

In the above, attention has been called to the great differ- 
ences in the spectra observed in sun, chromosphere and sun- 
spots. To find an explanation of the observed facts, it is 
necessary to look more closely into the newer developments 
regarding the structure of the atom. In Chapter XIV it was 
shown that many atomic characteristics are explained on the 
assumption that the electrons forming the more complex 


atoms are arranged in successive shells. The inert gases 
helium, etc. in column O of the periodic table, have atomic 
numbers of 2, 10, 18, 36, 54 and 86, and hence it has been 
assumed that the concentric shells contain 2, 8, 8, 18, 18 and 
32 electrons, respectively. 

Lithium of atomic number 3, and sodium of atomic num- 
ber n, are elements which have one external electron out- 
side of a complete shell. The process of removing an outer 
electron from an atom is called ionization. This may be 
accomplished by various electrical processes, some of which 
permit the measurement of the energy required in volts. As 
each electron carries a unit charge of negative electricity, 
an ionized atom, called an " ion," has a net positive charge 
of electricity. As we have seen, the spectrum of the ionized 
atom is radically different from that of the neutral atom. 
For reasons stated in the foregoing, the spectral lines be- 
longing to the ionized atoms are called " spark " or " en- 
hanced " lines. 

Not only can an atom be removed from an outer shell 
through ionization, but a given atom or ion can be excited. 
That is, the position of one or more of the electrons can be 
changed within the atom itself. To change from the normal 
state to an excited state requires energy. If this energy 
is absorbed by the atom, the spectrum refers to the dark line 
or absorption spectrum. When the change takes place in 
the reverse direction, from excited state to normal state, the 
same amount of energy is emitted by the atom and the 
spectrum is a bright-line or emission spectrum. As already 
stated in a previous chapter, the transference of energy 
from one state to another is always in definite amounts, or 
quanta, of energy. 

It has been found that the simple element hydrogen can 
exist in more than twenty excited states. More than one 
hundred different energy levels have already been discovered 
by spectroscopic means for the more complicated atom 
iron. The transference of energy from the excited to the 
normal state has been found to take place in the brief inter- 
val of time of approximately one-millionth of a second. 

Planck's quantum theory is justly regarded as one of the 


most important triumphs of the twentieth century. By 
means of it, we know that the frequency of the radiation 
emitted by the atom, and hence the wave-length, is exactly 
proportioned to the amount of energy radiated. 

In the case of hydrogen, we have seen that the wave- 
lengths of the 35 lines which appear in the chromosphere can 
be represented by the difference between two spectroscopic 
terms. With ordinary hydrogen and enhanced helium there 
is only one external electron. When there are two or more 
electrons, the mathematical computations become very com- 
plicated and as a result the exact calculations and predictions 
of spectral lines are impossible. However, there are sound 
reasons for believing that all electrons in the outer shells of 
atoms, no matter how complicated these atoms may be, 
obey rules, called quantum conditions, that are quite similar 
to those underlying the simple structures of hydrogen and 
enhanced helium. The changes in the outer electrons give 
rise to radiations which produce spectral lines. The energy 
of the electron depends on the relation of the size, shape and 
inclination of the orbits to each other, and hence with the 
more complicated atoms, the number of different energy 
states, or spectroscopic terms, is much greater than in the 
simple case of hydrogen. 

This is not the place to go into the technical details in- 
volved in the transference of energy from one spectroscopic 
state to another. For the more complicated atoms, it is im- 
possible to represent the orbits as was done so simply for 
hydrogen in Figure 4. However, by means of a diagram 
suggested by Bohr and Grotrian, the different energy states 
of an atom can be represented quite readily. Each state is 
represented by a dot. The distance below a heavy horizon- 
tal line is proportional to the amount of energy ( measured 
in volts) which must be communicated to the external elec- 
tron to pull it entirely away from the atom and thence to 
ionize the atom. The lowest dots in the diagram represent 
those of least energy while the heavy horizontal line repre- 
sents the ionized condition. 

The rules of the game ' laid down for the electrons to fol- 

1 See Russell, Dugan and Stewart, Astronomy, 548, 1927. 


low are not complicated. In the diagram, states correspond- 
ing to orbits of the same angular momentum plotted by 
dots in the same column thus s, p y d y f, g y h, etc., represent 
conditions of increasing angular momentum. When there is 
a transference of energy from one state to another, two of 
the dots are joined by a line. An atom in the is state (the 
lowest energy state in the s column) can change to any of the 
p states but not to the others; or an atom in the ip state can 
change to any s or d state, and so forth. 

In the sodium spectrum, for instance, when a small amount 
of heat is applied, as in the Bunsen burner, the only lines 
emitted are those confined to the two lowest levels s and p. 
The lines emitted correspond to transitions from the p states 
to the lowest s state. If white light is passed through sodium 
vapor these lines are absorbed. This series of lines is desig- 
nated by is mp. The level is is the normal state of 
the sodium atom in consequence of the lowest amount of 
energy. If the atom is excited and is in the ip state, it can 
return to the normal state only by the radiation of a quan- 
tum of energy corresponding to the transition 15 ip, 
which causes the appearance of the familiar D lines in the 

The transitions is mp are called the principal series. 
If the temperature is raised above that of the Bunsen burner, 
by the application of electricity, the atom can exist in higher 
energy states, such as 2^ and 2d. With increase of tempera- 
ture, the fraction of atoms in the ip state, though quite small, 
increases and hence the intensity of the lines become greater. 
At still higher temperatures, the fractions of the atoms in 
the 2$ and 2d states continually increase, with the consequent 
emission of lines represented by the transitions ip 2s 
and ip 2d, etc. 

If sodium is placed in the electric arc, then it is found that 
all the lines of the series is mp are sharp while the lines 
ip md are more fuzzy or diffuse. Hence the significance 
of the letters s and d. 

The amount of energy required to raise an atom from the 
normal state to any given excited state is called the excitation 
potential which is expressed in volts. A simple formula con- 


nects the energy 7, in volts, with the wave-length A, in ang- 
stroms as follows. 

V _i^45 

The p levels of Na, and also of all the alkali metals are 
double, and hence the transitions cause double lines in all 
the elements in column I of the periodic table just as with 
the familiar D lines of sodium. 

When there is more than one valence electron, the spectra 
are much more complex than that of sodium. There are 
then many different energy states and these may be single, 
double, triple and so forth up to eight-fold. The transitions 
from one energy level to another level give rise to numbers 
of lines known as multiplet groups. These may contain as 
many as fifteen members of varying intensities, some lines 
strong and some weak. 

All of these important spectroscopic developments have 
taken place in the past ten years. The chief workers in 
bringing order out of chaos, in the grouping of spectral lines 
into multiplets and deriving the excitation potentials have 
been A. Fowler, Russell, Saunders, Meggers and Miss Moore. 
Their investigations have been greatly helped through accu- 
rate wave-lengths furnished by King of Mt. Wilson and 
Meggers of the Bureau of Standards. 

The method of finding series relationships is somewhat 
similar to that used for the simple element hydrogen. In- 
stead of working with the observed wave-lengths it is found 
to be easier to work with frequencies, or the reciprocal of 
wave-lengths. Certain rules have been evolved to which the 
frequencies are subjected. Then in a manner somewhat simi- 
lar to that of working out a Chinese puzzle, relationships 
are gradually discovered between the lines of a certain 
element, and finally these lines can be grouped into 

Not only are the wave-lengths, or frequencies of the spec- 
tral lines found to obey the quantum conditions, but the 
intensities are also subject to laws. Since 1922 the prog- 
ress in this work has been very rapid until now all of the im- 


portant lines in the spectra of most of the elements have been 

In unravelling the puzzle, King's temperature classes have 
been of the greatest use. Lines of Class I appear at the low- 
est temperatures, that of the Bunsen flame. As the tem- 
perature is raised, first through the electric furnace and then 
at the higher and higher temperatures of the arc and spark, 
lines of Classes II, III, IV and V successively appear and the 
lines of the earlier classes become stronger and stronger. 
Class V lines appear in the arc only. Conversely, at lower 
and lower temperatures, all of the lines become progressively 
weaker and lines disappear in groups. The last lines to dis- 
appear, those that are visible even at the lowest tempera- 
tures, are called ultimate lines. These correspond to transi- 
tions to or from the lowest level. Lines from higher levels 
form subordinate series. 

With increase of temperature the energy to which the 
atoms are subjected increases. The rate at which atoms are 
ionized increases with the temperature. If the electrons 
liberated from the atom by ionization could be kept from 
combining with some other atom, then the process would go 
on until there were no more neutral atoms left. However, 
the liberated electron soon meets another ion and combines, 
again forming a neutral atom. Evidently the relative num- 
bers of neutral and ionized atoms depend on conditions. 
High temperature increases the rate at which atoms are 
ionized while on the other hand high density, where atoms are 
crowded together, increases the rate of recombination into 
neutral atoms. Hence it is evident that at a constant tem- 
perature, ionization increases in amount with decrease in 
density or pressure. At the same time it is evident that at 
the same temperature and pressure those elements can give 
up an external electron most freely when the energy required 
for this process, the ionization potential, is the least. Or in 
other words, the greater the ionization potential, the less the 
degree of ionization. 

As already stated, from a knowledge of heights to which 
vapors ascend in the chromosphere, Saha was able to show 
that close to the photosphere where the temperatures 


changed but little, increase of height meant decrease of 
density and pressure and a greater facility whereby atoms 
became ionized. Hence with the lines of the spectra grouped 
into multiplets for which excitation potentials are known, we 
are enabled to explain quite simply most of the problems 
connected with the chromosphere. 

Let us take as an example, the interesting element radium 
in order to see whether or not it is in the sun. Twenty years 
ago there was much discussion over the question, Dyson 
having found flash spectrum lines due to radium, while 
Evershed and Mitchell took the opposite stand. The strong- 
est lines of the principal series of enhanced radium are at 
3814 A and 4682 A, while three lines belonging to the diffuse 
series are at wave-lengths 3649 A, 4340 A and 4436 A. 
These are the strongest lines of Ra+, and if radium is in the 
chromosphere, we should unquestionably expect it to display 
its presence by these wave-lengths. As shown in Popular 
Astronomy, 21, 321, 1913, each of these five lines is already 
satisfactorily identified by coincidences with lines in Row- 
land's Tables, without invoking radium as a source. If 
therefore, we are to prove that radium is in the chromosphere, 
it will be possible only as the result of flash spectrum photo- 
graphs with much greater dispersion than those taken up to 
the present. 

The case of the element Mg is specially interesting for 
the reason that it is one of the alkali metals and is found in 
Column I of the periodic table immediately above the ele- 
ment Ca. The strongest lines of the solar and chromo- 
spheric spectra are the H and K lines of enhanced calcium. 
Unquestionably Mg is very abundant in the sun, but why 
then are there no very intense lines of ionized M g found in 
sun and chromosphere? The reason is not that the ioniza- 
tion potential is too high, for the values of Ca and Mg 
differ little, being 6.1 and 7.6 volts, respectively. In the 
chromosphere appear the three lines of the well-known 
6-group in the green, there is a triplet in the violet at 3829, 
3832 and 3838 A, all ascending to heights of 6000 or 7000 
km. Farther to the violet is another triplet at 3329, 3332 
3336 A. Each of the three triplets has an excitation 


potential of 2.70 volts. In addition there are many singlet 
M g lines in the chromosphere of the higher excitation poten- 
tial of 4.33 volts. These lines of Mg are all of them much 
more prominent than the corresponding lines in the spectrum 
of Ca. Hence it appears certain that Mg is a more promi- 
nent, and thus a more abundant element in the sun than Ca. 
But where are the ultimate lines of enhanced Mg corre- 
sponding to the H and K lines of enhanced Ca? The answer 
has already been given. They are far in the ultra-violet, at 
wave-lengths 2795 and 2802, and on account of the absorp- 
tion in our terrestrial atmosphere they cannot be photo- 
graphed either in the sun or in the chromosphere. The only 
line of Mg-\- of any importance in either solar or stellar 
spectra is the line at wave-length 4481 A. In the flash spec- 
trum this line exists but it is comparatively weak and is de- 
tected only to a height of 400 km. The reason is now clear. 
This line has a high excitation potential of 8.82 volts, it 
does not appear at all in the arc spectrum. Hence its pres- 
ence only in the spectra of the hotter stars is readily ex- 

Saha's theory thus interprets in a beautifully clear man- 
ner the systematic differences between the flash spectrum, 
the solar spectrum and the sun-spot spectrum. It goes 
much further, however, and furnishes the causes of the 
progression in type of the stars from the red stars of class M 
to the early types of B and O. Lockyer was the first to call 
attention to the change in the appearance of the lines H and 
K, very faint or even missing in late M stars, with a maxi- 
mum intensity in the solar, or Go stars, and becoming faint 
again in early B stars and disappearing in certain O stars. 
Lockyer 's interpretation, one of temperature only, was un- 
satisfactory. The hydrogen lines have their maximum at 
type Ao and are less intense in both the earlier and later 
types. The lines of neutral helium appear only in the stars 
of very early type, while the 4686 line and the Pickering 
series due to enhanced helium are found only in still earlier 
types. The conditions of appearance and disappearance of 
spectral lines due to ionization are calculable, and it has 
thus been possible to assign temperatures to stars of dif- 


ferent types which are in substantial agreement with those 
derived from other lines of research. All of the difficulties 
have not been entirely cleared away, but there has been a 
great step forward. 

The theory of ionization has shown the essential unity of 
astronomy, proving not only that the sun is a typical star 
but also that a study of the stars can shed much information 
on solar questions. The great problem of astronomy, the 
evolution of the stars and the structure of the universe, can 
find their complete solution only through an intimate study 
of the ultimate constitution of matter. The size of an atom 
can be deduced from observations of the gigantic stars. The 
structure of the atom and the theory of ionization are un- 
questionably among the most important problems of present- 
day physical science. 



THE conclusions drawn from eclipse spectra that 
strong lines originate at higher levels above the 
sun's surface than do lines less intense have very 
important consequences in the problem of the period of rota- 
tion of the sun. The determination of this rotation by spec- 
troscopic methods is one of the very greatest perplexity 
and still awaits an adequate solution. The information 
drawn from direct observation of sun-spots was positive 
enough as far as it went. At the equator on the sun, spots 
take about twenty-five days to make a complete circum- 
ference, while at latitudes north and south, the sun rotates 
more slowly , a day longer being required for a spot at 30 
than at the equator. On the whole, the information from 
spots furnishes contradictory conclusions since the indi- 
vidual spots have peculiar motions of their own which are 
not representative of the general surface of the sun. Few 
spots appear at distances greater than 40 from the sun's 
equator, and manifestly it is impossible to determine the 
law of rotation in high latitudes where observations on spots 
are impossible. 

With the application of the spectroscope and the photo- 
graphic plate to the problem, it was confidently expected that 
all difficulties would disappear, since the measurements were 
henceforth to be made on sharp and well-defined Fraunhofer 
lines. All that seemed necessary for a complete solution of 
the problem was to utilize a spectroscope of sufficiently high 
dispersion so that the shift in wave-length due to rotation 
could be measured with an adequate degree of precision. 
The great advantage of the spectroscopic method was that 
the observations might be carried out at any time without 
waiting for the appearance of a spot. A still greater ad- 



vantage, however, was that observations were not confined 
to the zones of sun-spots, but could be pushed even to the 
sun's poles. At the solar equator, the eastern limb advances 
at the rate of 2 .o km (1.2 miles ) per second, while the western 
limb recedes at the same rate. 

From observations of spots, several different empirical 
formulas have been devised to represent the rotation of the 
sun at different latitudes. Chief among them may be men- 
tioned those of Carrington and Faye. The latter has usually 
been regarded as the most satisfactory formula, and to it 
spectroscopic observations closely conform. Faye's equation 
may be readily adapted for spectroscopic work and takes the 
form l 

v + Vi = (a b sin 2 /) Cos / 

where v is the velocity in the line of sight deduced from the 
actual observations; v\ is a correction allowing for the orbital 
revolution of the earth so as to convert synodic periods to 
sidereal ; / is the latitude of the region on the sun under in- 
vestigation; and a and b are velocities measured in kilo- 
meters (or miles) per second. It has been assumed that a 
and b are constants, or, in other words, that the sun's rota- 
tion is not varying. If b is equal to zero, then the sun would 
rotate like a solid sphere with equal angular rotations in all 
solar latitudes. 

The first spectroscopic observations for determining solar 
rotation are more than forty years old, and were made by 
Duner of Upsala. Two lines were chosen in the red and 
their wave-lengths were compared with terrestrial lines which 
consequently show no rotational shift. The measures were 
carried to within fifteen degrees of the poles of the sun. Ob- 
servations have been in progress continuously for more than 
thirty years. The values for the first decade of intensive 
work have furnished J the basis for a detailed study of the 

After the expenditure of so much time and energy on 
this research it must be confessed that the results attained 

1 Newall, Monthly Notices, R. A. S., 82, JOT, 1921. 

2 St. John, Publ. Astron. Soc. Pacific, jo, .$19, 1918. 

These photographs taken in the red light of hydrogen show exquisite detail. 














are a grave disappointment. A glance at the velocities at the 
equator given in the table will reveal values ranging in size 
from 1.86 to 2 .08 km per sec. The measures were carried out 
very carefully, using every precaution to free them from er- 





km /sec 

No of 





2 08 



1900 5 



2 04 





Mt. Wilson 

2 06 





Mt. Wilson 

2 05 



1908 5 

Storey and Wilson 


2 08 




Plaskctt, J. S. 


2 01 




Plaskett, J. S. 


2 O2 



191 1 

De Lury 


I 97 






i 86 



igi i 

Plaskett, J S 


2 01 


4250 and 5600 




2 OO 




Evenhed and Kovds 


I 05 

3906 and 5624 


Plaskctt, II. II. 


i g8 

T 2 



St. John and Ware 

Alt Wilson 

i 94 




Plaskett, H II. 


1 95 




St. John and Ware 

Mt. Wilson 

i 04 




St John and Ware 

Mt. Wilson 

i 95 




rors. Each of the results in the above table is the mean of 
a very great number of observations on many different spec- 
trum lines (in one case numbering forty), and it would seem 
as a consequence, that the measures should have a high de- 
gree of precision with the final values entirely free from sys- 
tematic errors. Taken as a whole the measures seem to prove 
conclusively that during the two decades of observation the 
sun's speed of rotation has gradually diminished, showing a 
total change of five per cent. Before accepting this result as 
the truth, we must not forget the fact that the wave-lengths 
of lines in the spectrum have not the constant values originally 
supposed to exist; and to the many variations already known, 
eclipse observations add another, viz., that the stronger spec- 
trum lines occur at higher levels in the sun and must therefore 
display greater velocities of rotation. It is manifest that 
if the observations of one observer are to be compared with 
those of another, it will be necessary to subject the lines to 
a number of refinements as follows: ( i ) , Only lines with well- 


determined wave-lengths should be employed, they should 
not exhibit any " pole-effect," and to be representative lines 
they should not be " enhanced;" (2), lines of like Rowland 
intensities only should be used ; ( 3 ) , lines of the same region 
of the spectrum only should be utilized since we are enabled 
to see into the sun to greater depths in the violet than in the 
red, this being due to the scattering of light. These condi- 
tions will limit very materially the choice of available lines. 

Spectroscopic observations are ordinarily made at both 
ends of a solar diameter, the differential measures giving 
twice the value of the displacement. In order to eliminate 
any local effect at the limb due to the near presence of spots, 
faculae or filaments, it would be well to compare wave- 
lengths at the sun's edge with those at the center. The in- 
vestigations of St. John 1 show that the wave-lengths of lines 
at the center of the sun are " constant to a surprising degree 
of accuracy." Newall recommends the method developed 
at Cambridge, of making the measures of displacements be- 
tween east and west points on fixed chords, parallel to and 
equadistant from the projected axis of solar rotation, the 
east and west points being chosen in the same heliographic 
latitude, in either northern or southern hemisphere. This 
method is essentially one of comparing the law of solar 
rotation with the simple law of uniform rotation of a solid 
body. The method has some advantages over that ordi- 
narily in use, and some disadvantages. The peculiarity of 
this method is that the spectra are taken on the bright parts 
of the sun's surface and not near the fainter limb. The 
work of De Lury has shown that wave-lengths determined 
at the limb are subject to slight uncertainties due to the 
superposition of the sky spectrum. If the photographs are 
taken through thin haze, the wave-lengths may be displaced 
by amounts depending on the strength of the solar lines 

The discussions by Newall 2 and by Halm 3 of these spec- 
troscopic observations have led to interesting conclusions. 
The method usually followed by spectroscopists is to secure 

1 Mt. Wilson Contributions, No. 223, 1922. 3 Ibid., 82. 479, 1922. 

- Monthly Notices, R. A, S., 82, 101, 1921. 


observations on as many days, at as many different solar 
latitudes as possible. When sufficient material is obtained, 
the velocities in the line of sight being found, they are then 
subjected to a discussion by least squares in order to deter- 
mine the best possible values of the constants a and b in 
Faye's formula above. The different values of a, the velocity 
at the equator, obtained from various series of observations 
are listed in the table. 

It might be said, in passing, that if ever in the history 
of spectroscopic work positive and conclusive results were 
promised by any piece of astronomical research, the chances 
were very great that such results could be positively assured 
beforehand by the investigation on the rotation of the sun 
by spectroscopic methods. 

In the series investigated by the different observatories, 
various spectral lines and regions of wave-lengths were em- 
ployed. In spite of the possible sources of error to which the 
measures were subject, it would seem, if more than a dozen 
or fifteen lines were measured in a certain series, that any 
local peculiarities of a single line should be certainly aver- 
aged out. From the nature of the problem, it seemed alto- 
gether probable that lines of approximately the same aver- 
age Rowland intensity would necessarily be chosen by the 
different observers. Hence it might well have been expected 
that simultaneous observations of the sun made from differ- 
ent observatories would thoroughly agree in furnishing the 
same value of the solar rotation. How ideas have been 
changed in the past two decades regarding the constancy 
of wave-lengths! In comparing the longer series of ob- 
servations, we have the results tabulated below, where 
are given all of the observations secured since the year 
1906 as listed by Halm in Monthly Notices. In addition 
to the individual series secured at Edinburgh, Mt. Wilson 
and Ottawa, there is given the mean of all the values secured 
between the years 1901 and 1913, and also the results for 
the hydrogen line, Ha, obtained at Mt. Wilson. The veloci- 
ties are given in kilometers per second for every five degrees 
of heliographic latitude. 

The following facts should be noted regarding the quanti- 


ties in the table: ( i ), the values derived at Edinburgh agree 
perfectly with those taken almost simultaneously at Mt. 
Wilson; for among the seventeen separate linear velocities, 
only two differ as much as 0.02 km per second, showing a re- 
markable accord; (2), the three series at Ottawa are very 
consistent throughout all latitudes; (3), the three Mt. Wil- 
son series found in the third, fourth and fifth columns agree 




IQ06 S 

iqo6 4 

1907 o 

1008 s 


iyi2 s 

101 * S 


1 008 S 



Mt VV 

Mt W 

Mt \V 







2 03 

2 5 

2 07 

2 00 

2 01 

2 00 


2 04 



2 02 

2 03 

2 04 

2 04 




2 02 



I 98 


2 OO 

2 OO 







1 Q2 

I Q4 

I Q4 

i Q4 







I 80 

I 8 7 

r 80 

j 80 







I 7 8 

t 7 8 

i 77 

i 7 







t 09 

i 08 

i 00 

i <>5 







i 53 

i 57 

* 54 

i 54 





i Oi 


i 44 

i 43 

T 41 

i 40 


r 39 



1 ^0 


I 30 

1 29 

' 27 

' ~ 7 7 


i 24 


i 28 

T 37 


i 16 

i is 

1 U 

I I 2 

1 1 

I 1O 



I 22 


r 02 

I Ol 


o 97 

o 90 

o 05 

o 9S 

o 90 

1 08 


o 88 

o 87 

o 85 

o 81 

o 8[ 


o 80 

o 84 

o 9$ 


o 73 

o 73 

o 71 

o 07 


O 0* 

o 65 

o 70 

o 78 


o 60 

o 59 


o 53 

c 52 

o 48 

o 50 

o 55 

o 03 


o 45 

o 45 

o 43 

o 39 

o >S 


o 37 

o 41 

o 48 


o 30 

o 30 

o 28 

O 20 

o 24 

O 22 

o 24 

o 27 

o 34 

perfectly among themselves between latitudes o and 45, 
but show a progressive diminution of velocities at all latitudes 
poleward of 45 ; (4), the Mt. Wilson //a values are greater 
in amount than the Mt. Wilson results on the reversing layer 
for the year 19064, by an amount which is approximately 
constant and equal to 0.06 km per second ; ( 5 ) , the //a values 
exceed the velocities of the three Ottawa series by a constant 
which amounts to o.n km per second; (6), the time of ob- 
servations given in the Edinburgh and in the first Mt. Wil- 
son column (1906.4) agrees fairly closely with sun-spot 
maximum, while the mean of the three Ottawa series differs 
little from sun-spot minimum. 

On account of the, splendid accord between the observa- 


tions, we seem almost forced to conclude that the rotational 
velocity of the sun can be determined spectroscopically with 
great exactitude. The results tabulated above seem to 
prove most emphatically that at spot maximum the revers- 
ing layer revolves with a speed that is greater than the 
average, while at minimum of spots the rotational velocity 
is less than the average. This is the conclusion independ- 
ently reached by Newall and by Halm. 

A decade ago, at the 1922 meeting of the International 
Astronomical Union, the experts were forced to the con- 
clusion that, although there is such a beautiful accord be- 
tween the measures listed in the above table, it may not be 
impossible that some of the observations may have been 
affected by errors, either accidental or systematic in nature, 
and that in consequence, the difference in rotational values 
at maximum and minimum of spots may be illusory and not 
real. For the reasons already stated, systematic errors due 
to different spectral lines investigated may readily find 
place in any series of measurements, and the only conclusion 
to draw from the large differences between Mt. Wilson and 
Ottawa is that one or both of these series of measurements 
are unquestionably affected by systematic errors. In fact, 
investigations made at Mt. Wilson, in which simultaneous 
observations are made upon the two limbs and the center 
of the sun, indicate that temporary and local conditions fre- 
quently exist in the reversing layer which produce differ- 
ences of as much as ten per cent in the rotation values 
obtained by comparing east and west limbs directly. Such 
results make it obvious that misleading values may be actu- 
ally derived from a short series of observations; in fact it 
may be well to ask, along with the Chairman of the Com- 
mittee on Solar Rotation of the International Astronomical 
Union, " whether the solar rotation can be determined as 
definitely as has been thought." 

One obvious method of testing whether systematic errors 
affect the observational values and whether the rotation of 
the sun is different at sun-spot maximum from that at mini- 
mum is for one observatory with fine equipment to keep at 
the problem for a long stretch of years, taking great care to 


keep both methods and observers as constant as possible. 
Hence as early as 1914, a program of observations was inau- 
gurated at Mt. Wilson Observatory by St. John and Miss 
Ware. Subsequent to the values given in the table above, are 
the following linear equatorial velocities ' at the sun's equa- 
tor in units of kms per second : 

IQl6 igly 1918 1919 1920 1921 1922 192^ 1()24 T92cS 

i 91 I 94 1.95 i 90 I 90 i 91 i 91 i 89 i 91 i 89 

More recently, St. John 2 obtained values in 1929 and 1930. 
That in the latter year, amounting to 1.95, agreed in size with 
those of 1917 and 1918. Still more recently, Evershed* ob- 
tained the value 2.02 in 1931. It should be added, however, 
that Evershed made no distinction between high- and low- 
level lines and it is possible there may be a systematic error 
between his results and St. John's. 

Following the method of Adams, measures of rotation 
have been made on the high-level //a of the chromosphere. 
At Mt. Wilson in 1907, the value was 2.11 km per sec., in 
1919 this had changed to 2.00. However, in 1918, Ever- 
shed 's results gave 2.07 km per sec. while at Arcetri in 1929 
the value 2.07 was also obtained. Again we are faced with 
the conclusion that variations in instrumental equipment 
might readily permit different levels to be measured at two 
observatories with consequent systematic differences in 
velocities. The simplest method of explaining the large 
differences in the Mt. Wilson values is to assume that the 
measures of the bright line of Hex, in the two years were un- 
consciously referred to layers at different heights. 

With the splendid equipment at Kodaikanal, measures 
have been carried out on the dark Ha line of the Fraunhofer 
spectrum. The observations in the years 1918-22 were con- 
sistent and numerous and a value for the linear velocity at 
the equator was 2.03 km per sec. In the years 1923-29 the 
observations were not so numerous and they yielded a slightly 
increased value of 2.05. The weighted Kodaikanal value for 
the whole period 1918-29 is 2.03 km per sec. During the 

1 Mt. Wilson Obsy., Annual Report of Director, 127, 1928. 

2 International Research Council, Third Report, 121, 

3 Monthly Notices, R. A. S., 92, 105, 1931. 


same interval at Mt. Wilson the value from different lines 
was quite constant at 1.90 km per sec. On account of the 
higher level at which the Ha line has its origin, the results of 
the two great observatories are quite consistent in showing 
during a fifteen-year period, 1914-28, a solar rotation that 
appears to be nearly constant. 

In the latest annual report (1931) of the director of the 
Mt. Wilson Observatory, the intricate problem of solar 
rotation is carefully summed up in the following words: 
" It has become increasingly evident that the rotation of 
the sun as determined from observations of the reversing 
layer is not constant. Observations by Halm at Edinburgh 
and Adams at Mt. Wilson gave a velocity of 2.07 km per 
sec. and a minimum period of 24.75 days for 1905.5, while 
a smaller velocity was indicated by observations at Ottawa, 
Allegheny and Kodaikanal for the epoch of 1912, For the 
years 1919-29, the linear velocity averaged 1.90 km per sec., 
and the period two days longer than in 1905.5. Since 1928 
the linear velocity has slowly increased, while the period 
has decreased by 0.6 day. On the other hand, the velocity 
derived from long-lived spots has remained practically con- 
stant at 2. 02 km per sec. for the last thirty years. The 
lower reversing layer is 200 to 500 km above the photosphere 
(level of spots). In 1905.5 its linear velocity was 0.06 km 
per sec. greater than that of the photosphere, and in 1918, km per sec. less, thus indicating a streaming of the 
lower reversing layer relative to the photosphere from east 
to west during the spot cycle 1901-12, and from west to east 
during the following cycle. If connected with the direction 
of whirl in the spot-forming vortex, this systematic differ- 
ence would indicate a counter-clockwise rotation of the vor- 
tex in the northern hemisphere during the cycle 1901-13 
and a clockwise rotation during the following cycle. For 
the cycle beginning in 1923, the linear velocity of the re- 
versing layer might have been expected to increase, and 
apparently it has increased ; but whether it will continue to 
do so remains to be determined. A correlation between 
linear velocity of the reversing layer and direction of 
whirl in the spot-forming vortex would be of great inter- 


est in its bearing on the sign of the charge in the spot 

A summary l is given for other solar phenomena observed 
for the rotation of the sun. The number of degrees per day 
at or near the sun's equator shown from observations are: 
sun-spots, 14.39; faculae, 14.49; calcium dark filaments, 
14.45; reversing layer, 14.27; Ha (dark marking) 14.40; 
Ha (chromosphere, Adams), 15.00; K (prominences) 17.1. 
To find how the rotation periods change at different helio- 
graphic latitudes, one should also consult Abetti, Handbuch 
der Astrophysik, 4, 158, 1929. 

During the early years of the work at Mt. Wilson, the 
HOL line had shown such different rotational values from the 
other lines that it seemed advisable to extend the investi- 
gation to other solar phenomena. What values would be 
given by flocculi? To photograph in the red required special 
plates sensitive to this region. When the flocculi plates were 
taken, the results were a great surprise to Hale and Adams, 
for it was evident at a glance that the hydrogen flocculi dif- 
fered from those taken by means of the H- line of calcium 
in several remarkable features."' The photographs of hydro- 
gen flocculi already secured showed that they did not share 
the same retardation in rotation assumed by spots, faculae 
and calcium flocculi. But the striking differences between 
the details of the photographs of flocculi in the HOL light of 
hydrogen and in the H^ line of calcium, when examined side 
by side, showed that whereas most of the calcium flocculi are 
bright, those due to hydrogen are dark. Further differences 
were exhibited by the accentuated definiteness of structure 
of the hydrogen photographs which show details much 
smaller in size and with greater distinctness. The details 
brought forth by the Ha line were more marked than those 
obtained by the other hydrogen lines //(3, //y and 7/5, the 
reason for the difference being caused by the greater strength 
of the Ha line in chromosphere and prominces and the 
greater heights attained. A most unusual series of photo- 
graphs was secured by Hale on June 3, 1908, showing a dark 

1 Kodaikanal Bulletin, No, 113, 1931. 

2 Mt. Wihon Contributions, No. 26, 1908. 


hydrogen flocculus which was actually seen drawn into a 
sun-spot vortex, the remarkable change taking place within 
the brief space of time of ten minutes. The day previous, 
the location of the flocculus was evident from local whirls. 
On the day following the disappearance of the mass of cool 
hydrogen into the spot, photographs were secured showing 
eruptions in the neighborhood of the spot due to bright and 
hot hydrogen gas. Apparently, therefore, a sun-spot is a 
vortex somewhat resembling a terrestrial cyclone. Rela- 
tively cool matter in the gaseous form, floating high above 
the solar surface, is sucked into the vortex with a whirling 
motion. After sinking into the interior of the sun, the cool 
gas is heated, and later in the heated state makes its re- 
appearance outside of the limits of the spot, but in its imme- 
diate neighborhood. Meanwhile, the researches of J. J. 
Thomson and others had shown that electrified particles, 
both positively and negatively charged, must occur in vast 
numbers in a hot gaseous body like the sun. If there is a 
preponderance of positive or negative charges, their rapid 
rotation must give rise to a magnetic field. In other words, 
a sun-spot by its vortex motion appears to create its own 
magnetic field. 

The most promising method of attack appeared to be the 
Zeeman effect. When a luminous vapor is subjected to 
a magnetic field by being placed between the poles of a 
powerful magnet, an effect is produced on the lines of the 
spectrum. If the radiation is observed along the lines of 
force, the spectrum lines appear in most cases as doublets, 
having components circularly polarized in opposite direc- 
tions. It was found that different lines of the same element 
are affected to a different degree, and that the distance be- 
tween the components of a given doublet is directly propor- 
tional to the strength of the field. In a field of moderate 
strength the distance between the two components of a 
doublet may not result in complete separation, the lines be- 
ing merely widened, while in other lines, which are excep- 
tionally sensitive, the separation may be complete. As early 
as 1892, Young with the Princeton refractor, and later 
W. M. Mitchell, observed lines due to iron in the sun existing 


as single lines in the spectra of the photosphere but double 
in spot spectra. 

If, therefore, the observers at Mt. Wilson were to en- 
deavor to find a magnetic field in sun-spots, the line of at- 
tack was clearly outlined. First of all, there was necessary 
a large image of the sun, possible only by^the use of a tele- 
scope of great focal length so that the surface of the sun 
could be examined in regions surrounding the spot. The 
solar light must then be examined by a spectrograph of such 
great power that the separation of the doublets showing the 
Zeeman effect might be as great as possible. To interpret 
the measures of the solar photographs, investigations in the 
laboratory unquestionably would be necessary. Fortunately 
for the development of astronomical research, the resources 
of the Carnegie Institution were ample to provide, on the top 
of Mt. Wilson, the 6o-foot and the iso-foot tower telescopes 
with powerful spectrographs attached, and to equip in Pasa- 
dena a physical laboratory with the necessary forms of re- 
fined apparatus and Male's spectrohelioscope. As the result 
of twenty-five years of investigation, the scientific informa- 
tion garnered regarding spots and the general magnetic field 
of the sun has been very startling and most important. 

Observations are secured in the second order spectrum of 
the 75-foot spectograph attached to the iso-foot tower tele- 
scope. In front of the slit of the spectograph 1 is placed a 
Nicol prism in combination with a " quarter-wave plate " 
built up of mica strips 2 mm wide, mounted so that the prin- 
cipal sections of successive strips make an angle of 45 with 
the slit and 90 with each other. If the long axis of the 
Nicol is placed parallel to the slit, the mica strips will alter- 
nately extinguish the red and the violet members of a dou- 
blet, and the photograph will have a dentated appearance 
shown in the illustration facing page 332. It is hardly to be 
expected, however, that the extinction of either component 
will be complete. A partial extinction of the red or the vio- 
let component has the effect of shifting the maximum of 
brightness of the doublet towards the red or the violet of the 
average position, depending on whether the red or the violet 

1 Hale, Nature, 113, 105, 1924. 


component is the stronger. When the two components are 
not entirely separated, the lack of uniformity in the lines 
causes a displacement which depends on the relative strength 
of the two components. This unsymmetrical character of 
the lines renders the accurate measurement of the displace- 
ments one of the very greatest difficulty. In fact, even with 
the great scale of the Mt. Wilson photographs, reliable 
measures are possible only on a comparatively small num- 
ber of lines. 

The best line for the investigation of sun-spots is one in 
the red at 6173.553 A. This line is split into three compo- 
nents, the two outer ones being elliptically polarized in oppo- 
site directions, the inner component plane polarized. If the 
red component of the line is transmitted by the compound 
quarter-wave plate and the violet component extinguished, 
then the polarity of the spot is N, north-seeking or positive; 
while if the violet component is seen the polarity is S or 
negative. Since the distance between the components is 
proportional to the strength of the magnetic field, the direct 
measurement of a photograph by means of a micrometer 
readily furnishes the field-strength. In addition, the angle 
between the magnetic lines of force and the solar surface 
can be determined in all parts of a spot. 

Ever since the time of Galileo it has been noticed that 
spots have a tendency to appear in pairs. In fact, the ob- 
servations at Mt. Wilson show only that in ten per cent of all 
cases are the spots single or unipolar in appearance. Sixty 
per cent of spot groups are distinctly bipolar, while the bal- 
ance of thirty per cent show a tendency towards the bi- 
polar type, exhibiting calcium or hydrogen flocculi following 
or preceding a single spot. In many bipolar groups, one of 
the members, usually the preceding or western spot, is much 
larger than the other. In nearly all cases it has been found 
that the eastern member of a spot pair differs in polarity 
from the western member. When the group is formed of a 
number of spots small in size but without any dominant 
members, then the groups of spots at opposite ends of the 
stream are usually opposite in polarity. 

According to the opinion of Hale, a spot is a vortex in 


which the solar gases are expanded , and consequently cooled, 
by centrifugal action. If this cooling is sufficient, the gases 
appear darker over the vortex because of their decreased 
radiation and increased absorption at reduced temperatures. 
In the cooler regions, chemical compounds are formed, such 
as titanium oxide and certain hydrides, and their pr^ence 
is recognized by the existence of characteristic bands in sun- 
spot spectra. It is quite possible, however, that the cooling 
may not have advanced sufficiently far for the spot to become 
visible as a darkened area on the surface of the sun, but 
nonetheless it has been found that the magnetic effects per- 
sist even after the spot ceases to be visible and also appear 
before the darkening of the spot becomes sufficiently great 
to bring it into view. Thus have been discovered " invisible 
spots/' with the result that the life history of a spot is in- 
creased by lengthening out the times of observation both 
before and after the spot becomes visible to the eye. 

Ever since the beginning of the work at Mt. Wilson in the 
year 1908, many thousands of spots have been observed with 
great care. It has been found that all spots, independent of 
size, contain magnetic fields, but that the field-strengths up 
to a certain maximum increase with the diameter of the spot. 
Since the spots show themselves to be gigantic vortices, a 
natural question to ask is whether the directions of the 
whirls follow a law similar to cyclones and tornadoes on our 
earth which move in directions contra-clockwise in the north- 
ern hemisphere but right-handed in the southern hemisphere. 
Little is known regarding the sign of the charge dominant in 
sun-spots, but if it is assumed that these charges are the 
same in all solar vortices then the observed polarities will 
give the directions of the whirls. It has been found that 
preceding spots in the northern and southern hemispheres of 
the sun exhibit opposite polarities and consequently opposite 
directions of whirl. 

What is the law underlying the magnetic changes in sun- 
spots? From the time of the inception of the work at Mt. 
Wilson in June, 1908, to the minimum of spots in December, 
1912, there were a total of twenty-six groups observed. In 
seven of these groups that appeared in the northern hemi- 

?lf ; ^^^^fepll^ 

'i '" --'''" v.' n ' ' ., ; j - ny-to'*? 1 ** n H -V Vw= ^i i ; ' -\ J '""' "' '\t~' ~ "-' ''>'{-- ' - , "f fl " 

rlr --^^^V'*^:? '^ -; $& #' : _&,^ ^ : ;;>j?'_ ;J ' 

^- ' 


a, 6, ^ Photographic observations of a multipolar sun-spot group on August 8, 
1917 at Mt. Wilson Observatory; b and c photographed with Nicol and quarter- 
wave plate show the Zeeman effect with the iron triplet 6173 A. d An en- 
larged direct photograph of the same spot group on August 8, 1917 with the 
6o-foot tower telescope. (The black disk in corner shows the size of the Earth.) 





a * 



^ 31 


in +j 




sphere, the polarity of the preceding member of the pair, 
characterized by the violet member of the 6 1 73-line, was 
south-seeking or negative, and that of the following spot 
was north-seeking or positive. In the southern hemisphere, 
for seventeen groups the polarities were the reverse of those 
in northern latitudes, N or positive for the preceding mem- 
ber, and S or negative for the following member of the pair. 
Two southern spots out of the total of twenty-six observed 
showed opposite polarities to those expected, thereby fur- 
nishing exceptions to the rule. 

The spots which were observed were the dying members 
of the spot cycle and had an average latitude of 9. Would 








FIG. 7 Sun-Spot Polarities observed at Mt. Wilson between 1908 and 1925. 

The curves represent the change in latitudes of the spots; the preceding 
spot is shown on the right. 

the spots appearing in the new cycle, in high northern and 
southern latitudes continue the same persistence of polarities 
that the spots had already exhibited? There appeared no 
reason for a change, but when the observations were obtained 
they furnished a great surprise in showing a complete re- 
versal of polarities from those observed in the previous 

During the eleven-year period which closed in 1923 more 
than two thousand spot groups were observed at Mt. Wilson. 
Unipolar and bipolar groups, with only four per cent of ex- 
ceptions, continued to have the polarities reversed from that 
of the preceding cycle and showed south-seeking or negative 
polarities in the northern hemisphere, and north-seeking or 



positive polarities for the preceding spots of the southern 

Observations after the spot minimum in 1923 were anx- 
iously awaited to know whether a reversal would again take 
place with a resumption of polarities that had been the rule 
in the second preceding cycle. The answer was a very con- 
clusive one, the reversal which was now expected had actu- 
ally taken place. 

The law of sun-spot polarities is expressed in the two 
appended figures. At sun-spot minimum the new cycle origi- 
nates in high northern and southern heliographic latitudes. 

FIG. 8 Sun-Spot Polarities at Minimum of Solar Activity. 

Two zones in each hemisphere, in which the ^pots 

arc of opposite magnetic polarity, exist for about two 

years near spot minimum. 

As the cycle progresses towards maximum of spots they draw 
closer to the sun's equator but the polarity remains always 
the same. The high-latitude spots of each cycle have op- 
posite polarities from those of the preceding cycle. A strange 
condition appears near minimum of spots. In the northern 
(or southern) hemisphere two groups of spots differing a few 
degrees in latitude may exhibit opposite polarities, one po- 
larity due to a spot in the waning cycle, the other taking its 
sign from the cycle just beginning. Since the directions of the 
whirls of the hydrogen flocculi are not reversed, but, on the 
contrary, there is a change in the sign of polarity at each spot 


minimum, it is evident that the magnetic period of spots is 
22 years, or twice the n-year period on which the numbers 
of spots depend. 

It is evident that the general magnetic field of the sun 
must be much weaker than that exhibited in sun-spots on 
account of the fact that the sun as a whole rotates much 
more slowly than the vortex whirlpools of the spots. More- 
over, in attempting to secure the magnetic field of the sun as 
free as possible from any local effect of sun-spots, it is obvi- 
ously necessary to obtain photographs when the sun is en- 
tirely free from active spots. Several series of photographs 
taken under satisfactory conditions have been secured. The 
difficulty of measuring these photographs, and of securing 
these measures free from personal and systematic errors, is 
patent to anyone who has ever engaged in any astronomical 
measurements requiring the setting on two close components 
of a double. Moreover, the size of the quantity to be meas- 
ured is only o.ooi A. As a matter of fact, van Maanen was 
the only one of the five Mt. Wilson measurers that engaged 
in the work whose results had a satisfactory degree of con- 
sistency. On account of the difficulties in measuring, lines 
having an intensity greater than 5 in the solar spectrum and 
weaker than o had to be excluded. 

The general summary of the results seems to prove con- 
clusively that a general magnetic field exists in the sun, and 
that in consequence the sun behaves approximately as a 
uniformly magnetized sphere, with the magnetic axis only 
slightly inclined to the solar axis of rotation, and with a 
polarity corresponding to that of the earth. Forty-six spec- 
tral lines were investigated, and of these, 30 lines due to e, 
Cr, Ni, V and Ti show displacements. The strength of the 
magnetic field determined for each line showed a correlation 
between field-strength and line intensity, the stronger fields 
being connected with lines of smallest intensity. Eclipse spec- 
tra reveal the information that the weakest solar lines origi- 
nate at the lowest depths. Since it was possible to measure, 
for the determination of the strength of field, lines of a solar 
intensity 5 or less, it is manifest that lines only in a very shal- 
low layer, less than 450 km in depth, can show the influence of 


the sun's magnetic field large enough to be detected by the 
present method of attack. The period of rotation of the 
sun's magnetic axis was found to be 31.52 days. No explana- 
tion can be given for this peculiar value of rotation which 
differs in such marked degree from the equatorial value de- 
rived from sun-spots. In fact, no adequate explanation is 
yet forthcoming to explain the cause of the sun's magnetism. 

It is now generally conceded by all scientists that the sun 
is continually sending off a vast stream of negatively charged 
particles, or electrons. It is also generally conceded that 
the presence of spots on the sun is evidence of great solar 
activity. If the spot is large, and the activity consequently 
great, the streams of electrons directed toward the earth 
may reach our upper atmosphere in vast numbers. In the 
rarefied conditions that must exist in the upper atmosphere, 
the air becomes ionized. This ionization causes the electro- 
magnetic display known as northern lights, or aurora bore- 
alis. As a result of the electrification of the atmosphere, 
currents are induced in the earth, which may at times be so 
strong as to seriously interfere with the sending of messages 
over the telegraph lines or submarine cables. As a further 
manifestation of the currents flowing through the earth, the 
navigator's compass may be affected with the result that the 
north end of the needle may point several degrees from 
the magnetic north. This leads to a so-called " magnetic 
storm." The display of a brilliant aurora is very frequently 
accompanied (and caused) by a spot central on the face of the 
sun. On May 13, 1921, the spots near the sun's center 
caused an unusually gorgeous display of aurora visible prac- 
tically all round the world. The telegraph lines were seri- 
ously hampered in the sending of messages. While the 
aurora was at its height, one of the Atlantic cables connect- 
ing Europe and America was burnt out. Whether this was 
caused by the earth's currents, or was but a strange coinci- 
dence, has never been fully determined. 

Theoretically, sun-spots should possess an electric as well 
as a magnetic field. Although observations have been made 
at Mt. Wilson and elsewhere, no conclusive evidences of a 
Stark effect have been recorded. 


In an important paper presented * at the National Acad- 
emy of Sciences, Adams and Nicholson discuss The Nature 
of the Solar Cycle in the following manner. " The eleven- 
year period may vary between nine and fourteen years and 
the amplitude of the cycle by about 50 per cent of its average 
value. Period and amplitude are apparently unrelated. It 
seems probable that both the quantity and quality of solar 
radiation vary during this cycle and many attempts have 
been made to correlate terrestrial phenomena with sun-spots. 
Definite correlations have been found with the variations in 
terrestrial magnetism and its related phenomena. There is 
evidence of a slight correlation between sun-spots and atmos- 
pheric temperature in certain regions on the earth and with 
other factors of weather and climate for limited regions and 
for limited time intervals. These correlations are so uncer- 
tain that, in the majority of cases at least, predictions on 
them have very little weight." 

1 Science, 75, 594, 1932. 



NEW information gleaned in the past decade about the 
structure of the atom has revolutionized the method 
of attacking solar problems. Practically all of the 
prominent lines in the spectrum of the sun have been assigned 
to multiplets with known excitation potentials, the arbitrary 
intensity scale of Rowland has been submitted to calibration 
tests which have revealed that the intensities depend on the 
number of atoms engaged in the formation of the spectral 
lines. From the weakest lines perceptible in the solar spec- 
trum of intensity - 3 ( or oooo ) to the strongest Fc lines at 
wave-length 3720 and 3735 of intensity 40, the number of 
atoms involved increases about one million times. 

A person having no knowledge of the theory underlying 
multiplet groups would not advance very far in the practical 
operation of correlating heights in the chromosphere with in- 
tensities either in sun or chromosphere before the fact would 
be forced upon his attention that generally the lines of great- 
est intensity reach the greatest heights, and moreover the in- 
tensities and heights for any element are greatest for the 
multiplets of lowest excitation potential. The best element 
for a study of the explanation of these correlations is unques- 
tionably neutral Fe. This element has been assiduously ob- 
served in the laboratory and very exact wave-lengths are 
known. Fe is very rich in lines which have been grouped into 
multiplets with a wide range of excitation potentials. Some 
of the more important results from the detailed study x of this 
element are given herewith. The lowest atomic level of neu- 
tral Fe is the a G D level with average excitation potential of 
0.07 volts, the next lowest level is a~'F, with excitation poten- 
tial of 0.92 volts. The table gives the average heights in the 

1 Publ. Leander McCormick Ob\y. f 5, 114; Astroph. Jour., 72, 146, 1930. 



chromosphere attained by lines of different Rowland inten- 
sities, with the whole material arranged according to in- 
creasing excitation potentials. A total of 789 lines is 


Excitation Potent a 

20 jo 











0.07 . . 


I 4 




1 100 










0.92 . 


I 7 














/i 650 

I 2 










2 18-249 . . 

/ 775 

\ 4 



1 7 






2 50-2 99 

/ 500 
1 i 

i 5 













/ 542 














i "I 









4 00-4 49 ... 

/ 4 

1 i 











/ 4O 

I 4 







* The upper quantity give* the height in kilometers, the lower gives the number of spectral linea involved 

The following facts should be noted: (i) The strongest 
lines of neutral Fe in the sun belong to multiplets of lowest 
excitation potential 0.07 volts. (2 ) With increase of excita- 
tion potential, the maximum intensity of the lines in the mul- 
tiplets steadily decreases. ( 3 ) For multiplets of any given 
excitation potential, there is a close correlation between in- 
tensities and heights. (4) For any given Rowland intensity, 
such as 6, the heights diminish (vertically in the column) as 
the excitation potentials are increased. To know the height 
corresponding to a line of intensity 6, the value of the excita- 
tion potential is necessary. 

If the material from the flash spectrum on neutral Fe is 
divided at wave-length 4900 A into two groups, we have the 



information in the following table. The first two horizontal 
lines of the table give the average heights and average ex- 
citation potential of the lines to the violet of 4900 A ar- 
ranged according to their Rowland intensities. The next 
two lines give similar data from the lines to the red of 
4900 A; and next comes the average for all lines of all wave- 
lengths. The two last lines in the table give the radial mo- 
tion in the penumbra of sun-spots called the Evershed effect, 
measured by St. John at Mt. Wilson, the first line giving the 
value in angstroms and the second line in kilometers per 










/ 3>I 







Oo >> 




i 2 20 

2 95 

I oi 

2 70 


2 54 

2 29 

2 2J 

i 82 

r 90 


1 287 





40 s 






I 3 bo 

3 44 



i U 

3 10 

2 57 

2 OS 

I 82 

2 }2 


I 290 

Hg , 

30 1 






08 1 



\ 3 H 


* *5 

2 90 


2 77 

2 39 

2 ^0 

i 82 
















i 5 

I 4 

I 2 









O 1 

IS 18 


O 1 




o o 

Early in the year 1909, Evershed announced a remarkable 
discovery of far-reaching importance in his observations of 
the displacement of Fraunhof er lines in the penumbra of sun- 
spots. With the slit of his spectograph placed across the 
spot, he found that the wave-lengths of lines in the penumbra 
of the spots were different from the values at the center of the 
sun. The displacements, which affected practically all of the 
lines of the reversing layer, were not constant but differed in 
amount depending on the intensity of the lines investigated. 
The shift was greater for the weaker lines of the spectrum 
than it was for the stronger lines. Evershed advanced the 
hypothesis that the observed displacements are the result of 
the Doppler effect, and that in consequence, the gases of the 
reversing layer are in radial motion tangential to the solar 
surface. 1 

1 Kodiakanal Observatory Bulletin, No, XV: Kodiakanal Observatory Memoirs. 
r, Pt. i. 







Following the announcement of this important discovery, 
St. John began an extended series of investigations into the 
subject, the results of the observations being published in the 
Contributions from the Mount Wilson Observatory, Nos. 69, 
74, 88, 348 and 390. The observations were carried out with 
the 6o-foot tower telescope, with the image of the sun 170 
mm in diameter, the penumbra of the spots investigated aver- 
aging 3.0 mm in diameter. The plates in the violet and green 
were taken in the third order spectrum, and those in the yel- 
low and red in the second order. For the two cases, the dis- 
persion was i mm = 0.56 A and i mm = 0.86 A, respectively. 
Measures were carried out on 506 lines, some of the lines be- 
ing measured on no less than thirty plates. 

The Mt. Wilson measurements, so carefully made by Miss 
Ware, abundantly verified Evershed's conclusions that the 
displacements are caused by movements of the solar vapors 
tangential to the solar surface and radial to the axis of the 
spot. These motions are none other than the actual flow of 
the material of the reversing layer out of the spots and of 
the matter forming the chromosphere into the spot vortex. 

The layers closest to the sun's surface have a motion of 
translation out of the spot at the rate of two kilometers per 
second, and this motion becomes less and less at greater and 
greater elevations until, at a height of about two thousand 
kilometers, the motion outward of gases from the sun-spot 
ceases. What happens to the vapors above this level? The 
information furnished by the investigations of St. John is very 
definite and apparently admits of no contradiction. From 
measurements on several elements, it was found that above 
this level of inversion, the gases of the chromosphere take a 
motion carrying them in to the spot. As greater and greater 
elevations are reached these movements increase in amount. 
At the maximum heights reached by lines of the chromo- 
sphere, which are attained only by the H and K lines of the 
element Ca, there is a movement of the calcium vapor into 
the spot at a speed of 3.8 km per second corresponding to a 
displacement measured by St. John of -0.063 angstroms. 

In the following table there is collected the available in- 



formation from the measures of heights derived from the 
lines in the flash spectrum and of displacements due to the 
Evershed effect from 506 lines in the penumbra of sun-spots. 
The + symbol affixed to an element, as Ca +, signifies that 
the line in the spectrum is enhanced and that it takes its origin 
from the ionized atom. The conclusions of St. John (loc. 
cit.} are, " In the observation of these velocities we have a 
methol of sounding the solar atmosphere and of allocating 
the relative levels of the lines." 


Currents measured in Kms/sec 

Heights in Kms 
from Flash Spectra 





Ca+ (II and K) 

3 8 inwards 

o 5 downwards 





77/3, 77*, 77f 


77 7 , 776 

i i 





o 7 ' 


77+,Sr+, k Vr+ 

o 5 


Ca (4^27) 

O 2 * 


Al (15-20) 

O O 

i ,500 

Fe (15-40) 

o i outwards 

o 3 downwards 


Fe do) 



Fe (8), TV, Sc, V 

o 6 



Fe (4), 0, Sr, V 

I 2 


Fe (2), A7, Co, Mn 

i 5 


Fe (i) 

i 7 

o 3 upwards 


Fe (oo) 

2 O 

It is evident, from a glance at the above table, that the 
enhanced lines not only extend to greater elevations than 
do the unenhanced lines of the same element, but that the 
Evershed effect for them is more pronounced as well. 

Eclipse spectra, taken in connection with the Evershed 
displacements, can shed some light on the problem of sun- 
spots, not only regarding the time-honored question of the 
heights above the photosphere at which the spots themselves 
originate but also regarding the magnificent work of Hale on 
spot vortices. After long years of uncertainty regarding the 
effective temperature of spots, whether they are hotter or 
cooler than the photosphere, we have finally come to the con- 


elusion that the spots are relatively cooler. This fact first 
became known through the discovery by Fowler of bands in 
the red part of the spectrum due to titanium oxide. Bands 
and flutings of hydride compounds have likewise been found 
in parts of the spectrum other than the red. There seems 
only one explanation possible, which is that the compounds 
can exist only at the cooler temperature of the spots, and that 
at the higher temperatures of the photosphere the com- 
pounds are dissociated and their bands cease to show. All of 
the researches concerning spots confirm this view of cooler 
temperatures, the low-temperature furnace lines being 
strengthened and the high temperature enhanced lines being 
weakened in spots. 

If we are to accept the explanation of the Evershed dis- 
placements that they are a Doppler effect which differs for 
lines of various solar intensities, which in turn depend on the 
depths at which these lines originate, then we are forced to 
the conclusion that if there are differences in level between 
the spots and the photosphere these differences cannot be as 
much as fifty kilometers. 

The motions in the line of sight of the gases in the penum- 
bral regions of sun-spots can be measured with high precision 
on account of the great dispersion employed at Mt. Wilson. 
When the measures of the Evershed effect from the individual 
Fe lines are combined into averages, and the spectral lines are 
then arranged in various manners according to Rowland in- 
tensities, heights from the flash spectrum, excitation poten- 
tials, and spectral regions, conclusive evidence is found that 
the size of the Evershed effect does depend primarily on the 
heights. On the assumption, now thoroughly well founded, 
that a spot is a vortex on the sun, it appears abundantly veri- 
fied that the indirect determination of relative heights from 
the Evershed effect tell a story consistent with the direct 
measurement of heights from the flash spectrum. 

By referring to the above table, these interesting facts are 
observed : ( i ) To get the same average height in the red as 
to the violet of 4900 A, it is necessary to add two units of in- 
tensity to the right in the table, or 2 units of greater strength 
in the red. ( 2 ) In quite similar fashion, to reach the same ex- 


citation potential in the red as in the violet, it is necessary to 
go to lines of 2 Rowland units of greater intensity in red than 
in violet. In addition (3) St. John has found that to get the 
same measured value of the Evershed effect in the red as in 
the violet, it is necessary to go to 2 units of greater intensity 
in the red. 

From the calibration of Rowland's intensity scale, it was 
found that the number of atoms involved in the formation of 
a line of Rowland intensity n in the violet is approximately 
the same as are engaged in the production of a line of Row- 
land intensity n -f 2 in the red. Hence we see that a spectral 
line produced by a given number of atoms has the same Ever- 
shed effect, the same height in the flash spectrum and the 
same average excitation potential throughout the spectrum. 
It is evident that there is one underlying cause, and one only, 
needed to explain these interesting correlations, namely, the 
number of atoms involved. But how has this information so 
vitally necessary for explaining many of the problems con- 
nected with solar radiation been determined? 

After arranging the lines of different spectra in multiple! 
groups, it was found by a number of investigators 1 working 
independently, and almost simultaneously, that simple for- 
mulae give the relative intensities of lines belonging to a given 
multiplet. The formulae are based on the correspondence 
principle, and give the transition probabilities as a function 
of the quantum numbers. These probabilities when multi- 
plied by p 4 , v being the frequency, are proportional to the en- 
ergy emitted in the separate lines. Measures show that these 
formulae are only approximate, but on the whole they repre- 
sent the intensities with fair degree of accuracy. According 
to Ornstein, 2 the quantities given by the formulae represent 
the relative numbers of " fictitious resonators," or the relative 
numbers of atoms involved in the production of the lines of 
the multiplets. 

Let us first see, therefore, how the simple rules provided 
by the formulae can represent different arbitrary estimates of 

1 Kronig, Zeits. fur Physik, 31, 885, 1925; Sommerfeld and Honl, Sitz. der 
Preuss, Akad. der Wiss., p. 141, 1925; Russell, Proc. Nat. A cad. of Sciences, u, 314, 

2 Zeits. fur Physik, 40, 412, 1926. 


intensities. Take first the estimates by A. S. King, many 
thousands in numbers of the spectra of the lines in the labora- 
tory. Russell (loc. cit.) has found that King's scale, which 
was intended to represent the actual intensities of the lines, is 
a remarkably homogeneous scale. These Mt. Wilson inten- 
sities in fact are very nearly proportional to the square roots 
of the relative numbers of atoms predicted by the theory of 
multiplets. " The agreement is so close that these estimates, 
especially when averages for several multiplets are available, 
are clearly almost as valuable as actual measures, when once 
the significance of the empirical scale has been found." 

Another comparison of even greater importance is the cali- 
bration of Rowland's scale of intensities for the solar spec- 
trum. Russell, Adams and Miss Moore by means of the 
intensities of 1288 lines grouped in 288 multiplets have inves- 
tigated l Rowland's scale in the solar spectrum. The relative 
intensities in multiplets, based on the correspondence princi- 
ple, can be assumed to be proportional to the numbers of ac- 
tive atoms producing the various lines. These results have 
most important consequences. 

By the observed contours of a number of typical resonance 
lines of different elements, Unsold has determined J the num- 
ber of atoms above i sq. cm of the sun's surface which are 
concerned in the production of the lines. Although many 
theoretical and practical difficulties have still to be sur- 
mounted and much more observational work must be accom- 
plished on the contours of lines, nevertheless we have as a 
result what is perhaps the most important astrophysical pub- 
lication in recent years. This comes from the skillful hands 
of Russell J entitled " On the Composition of the Sun's At- 
mosphere." Unfortunately space will not permit a detailed 
discussion to be given here. 

One very important result was the derivation of the energy 
of binding of an electron in different quantum states by neu- 
tral and ionized atoms. These were given for various ele- 
ments in tabular form together with tables of ionization po- 
tentials. These principles were first applied to enhanced 

1 Astrophysical Journal, 68 , i, 1926. : * Astrophysical Journal, 70, u, 1929. 

- Zeits. fur Physik, 46, 765, 1928. 


lines. Russell finds that thirteen elements have their most 
persistent lines in the region of the spectrum accessible to 
photographic observation. These elements are Be y Ca, Sc, 
Ti, V, Sr, Y, Zr y Cb, Ba, La, Hf, and Ra; all but the last, 
radium, have enhanced lines in the sun. Other important ele- 
ments like Cr, Mn, Fe, Co, Ni and Mo have their most persist- 
ent lines in the inaccessible region in the sun to the violet of 
wave-length 3000, but other multiplets arising from low 
energy-levels are accessible for investigation. Fortunately 
with few exceptions enhanced lines are found in the solar 
spectrum for all those elements which have lines of low ex- 
citation potential in the accessible region of the spectrum. 
For reasons stated in foregoing chapters, these elements have 
their lines strengthened in the chromosphere and hence are 
prominent in the flash spectrum. 

On the contrary, the principal factor which is unfavorable 
to the appearance of a spectral line in the sun is a high excita- 
tion potential. As a matter of fact, there are comparatively 
few lines in the solar spectrum with excitation potential 
greater than 5 volts, the only lines of great strength being due 
to hydrogen. (In the flash spectrum, however, is found the 
He 4~ line at 4686 A, with excitation potential of 48 volts.) 

Hence the excitation potentials for the strongest lines in 
the visible part of the spectrum give the means of explaining 
the presence or absence of lines of different elements in the 
sun. If the most persistent lines are not observable in the 
solar spectrum, as with the element Mg, the method can give 
only a minimum value of the abundance. 

These theoretical considerations depend on a knowledge of 
temperature and pressure in the sun. According to Edding- 
ton/ the temperature of the reversing layer doubtless is great- 
est closest to the photosphere and may be assumed to be 
5730 absolute. Russell finds a pressure of 3 x 10 f; atmos- 
pheres, in good agreement with the conclusions of other 

The total quantity of the metallic elements in the solar at- 
mosphere seems to be known with some accuracy. The prob- 
lem of the non-metals is more difficult, mainly on account of 

1 The Internal Constitution of the Stars, p. 332, 1926. 


less complete observational material. The metals provide all 
but one-half of one per cent of all the ions and electrons. If 
the correction for departure from thermodynamic equilibrium 
should be disregarded, the calculated abundance of hydrogen 
would be increased thirty-fold. 

Assuming the abundance of hydrogen found by Menzel 
from the flash spectrum (and which the present writer re- 
gards to be subject to large systematic errors), then Russell 
finds the probable constitution of the sun's atmosphere to be: 

Element By Volume By Weight 

Hydrogen 60 parts 60 

Helium 2? 8? 

Oxygen 2 32 

Metals i 32 

Free Electrons 0.8 o 

Total 65.8 132 

Tables are given for the relative abundance of fifty-six 
elements and six compounds. Six of the metallic elements, 
Na, Mg, Si, K, Ca and Fe contribute 95 per cent of the total 
mass of all the metals. 

Comparisons made with Miss Payne * show an excellent 
agreement in abundances for eighteen of the most important 
elements, except hydrogen and also the element K (prob- 
ably because Miss Payne's value for the latter depends on 
only two lines). Comparisons also made with the abun- 
dance of elements in the outer ten miles of the earth's crust, 
including the ocean and atmosphere, and in stony meteorites, 
show a good agreement throughout (better with meteorites 
than with sun), except again hydrogen. The difference for 
H is enormous, part of which is probably real and part is 
due to the uncertainties of the calculation. 

Russell calls attention to three puzzles still outstanding: 
( i ) the calculated abundance of hydrogen in the sun's atmos- 
phere is almost incredibly great; (2) the electron pressures 
calculated from the degree of ionization and from the num- 
bers of metallic atoms and ions are discordant; and (3) the 

1 Harvard Observatory Bulletin, No. 835, 1926. 



calculated rate of increase of density with depth in the revers- 
ing layer is much more rapid than indicated by observations 
of the flash spectrum. 

With the important problems involving numbers of atoms 
so clearly stated by Russell and many other investigators, we 
are now in a position to take up anew the question of what 
may be expected within a multiplet of any element. 

THE Fe MULTIPLEX a 5 F y 5 !) 



I 1 A 



( 320 



1 25 









01 6 











\ 1500 



1 ooo 






f 3834 







\ I2OO 






















Intensity in sun 


Do . ... 

Height in kilometers 


Evershed effect in sun-spots 
Sun minus vacuum arc in angstroms 


In important articles, St. John ' and Burns ~ have discussed 
their own particular points of view regarding the observa- 
tional details within a given multiplet. For the purpose of 
illustrating the " unit nature " of multiplets, both have util- 
ized the same Fe multiplet, a r 'F y 5 D, with excitation po- 
tential 0.92 volts. After a thorough study of the whole 
problem, by comparing this njultiplet with others, and then 

1 Mt. Wilson Contributions, 389, 300, 1930. 

2 Journal of the Optical Society of America, 20, 212, 1930. 




o o 


O M 

w . 

dl fO 








5! b 

> 2 

3 2 

B fe 








o - 


5) cd 

CL l-i 

3 1 

w S 

B 5 




smoothing out accidental errors and getting rid of the effect 
of blends, the appended table is adopted by the author as 
his idea of this composite, or ideal, multiplet. The top num- 
ber in each case is the wave-length, below this is the intensity, 
then come in order for each line, the height in kilometers 
from the flash spectrum, the Evershed effect measured in 
angstroms, and finally, the difference in wave-length between 
the value in the sun minus that of the vacuum arc. The Ein- 
stein, or relativity, shift to the red amounts to 0.0082 A, or 2 1 
parts in ten-million in wave-length. 

In this multiplet, the numbers of " fictitious resonators " 
effective in forming the different spectral lines vary greatly. 
According to Russell ( loc. cit., p. 326) more than one hundred 
times as many atoms are active in producing the strongest 
line of the multiplet at 3820 A as go to form the weakest line 
at 3940 A. Hence, within a single multiplet one should expect 
that the heights found directly from the flash spectrum, or 
indirectly from sun-spots, would not be constant but would be 
greatest for the largest numbers of atoms. 

St. John and Burns differ radically in the interpretation 
of the systematic differences between the observations for 
lines of different intensities in the table. The former takes 
the ground that the relativity shift to the red in the sun has 
been confirmed, while the latter is still unconvinced. They 
agree, however, that the difference sun minus vacuum arc 
should be a constant. Burns finds (as shown in the table) 
that the more intense lines have a greater red displacement 
than the less intense lines in the same multiplet but this is di- 
rectly contrary to St. John's deductions. Burns voices his 
opinion in the following words: " It seems paradoxical to as- 
sume as a possibility that a very strong line of a multiplet can 
originate at a high solar level, where St. John postulates the 
descending movement, while a weak line of the same mul- 
tiplet originates only at a lower solar level in the ascending 
vapor. " 

The present writer is inclined to agree with St. John and 
therefore to disagree with Burns in the question of the rela- 
tivity shift in the sun; he agrees with Burns and disagrees 
with St. John as to the unequal displacement for different 


lines in a multiplet; but he disagrees with both St. John 
and Burns in the interpretation of the data observed within 
single multiplets. 

Two strong multiplets of Fe are specially valuable for pur- 
poses of comparison, namely , a 'D z D () , with average ex- 
citation potential 0.07 volts, and a r 'F y r 'D, with excitation 
potential 0.92 (the composite multiplet given in the table). 
These two multiplets have the same maximum Rowland in- 
tensity of 20 in each case, and nearly the same minimum 
intensity. Moreover, the two multiplets cover approxi- 
mately the same spectral region, from 3820 to 3940 A, and 
hence no systematic effects depending on wave-length need 
be considered. Although the Rowland intensities average 
the same in the two multiplets, there are pronounced differ- 
ences in the other observed quantities. The multiplet of 
low excitation potential 0.07 volts compared with that of the 
higher value 0.92 shows the following: ( i ) greater intensities 
in the flash spectrum, (2) greater heights and (3) greater 
values of the wave-length difference sun minus vacuum arc. 
Each multiplet separately for ( i ) , ( 2 ) and ( 3 ) show higher 
values on the main diagonal than on the side diagonals. For 
a height of 2000 km from the multiplet with excitation po- 
tential 0.07, the difference sun minus vacuum arc is -f- 0-0135 
angstroms which, corrected for the relative shift 0.0082 A, 
gives a residual difference + 0.0053 angstroms correspond- 
ing to a downward motion in the sun's atmosphere of 0.4 km 
per second. 

It must be emphasized again and again that the " ef- 
fective height " above the photosphere at which a line in the 
Fraunhofer spectrum is formed, the heights from eclipse 
spectra and the heights from the Evershed effect, do not and 
cannot refer to the maximum heights to which atoms are 
ejected by solar activity. All three heights are derived from 
photographic plates, all depend on the effects produced by 
the atoms on the photograph. As already stated in the fore- 
going, the atoms must be in sufficient numbers, or have a 
concentration adequate to leave a trace of their action on the 
photographic plate. It is not impossible that under average 
conditions of solar activity, all Fe atoms no matter what are 


their atomic levels or whether they are neutral or enhanced 
reach about the same maximum heights in the chromosphere. 
If the solar activity becomes greater, as in the course of 
the sun-spot cycle, or locally as in prominences, the heights 
become greater. 

The conditions under which atoms in the chromosphere 
radiate to form the flash spectrum (or absorb radiation in 
the dark-line spectrum) are becoming well understood. 
They depend on temperature, pressure and concentration of 
atoms. For a single multiplet, it is obviously necessary to 
assume that all of the spectral lines are formed under nearly 
identical conditions of temperature and pressure and that all 
the atoms reach the same maximum heights. The lines dif- 
fer greatly the one from the other in the number of atoms in- 
volved in the emission (or absorption) of radiation. Let us 
now confine our attention to two lines from a multiplet, the 
one produced by one hundred times as many atoms as the 
other. Consider for each line a cylinder of unit cross sec- 
tion. The base of each cylinder we shall call the photosphere, 
each axis being perpendicular to the solar surface. The top 
of each cylinder is at an equal, but at a very great distance 
above the photosphere. The atoms are most concentrated 
near the photosphere. Looking down into the axis of each 
cylinder, that is, observing a spectral line at the center of 
the sun's disk under the ordinary conditions of the Fraun- 
hofer spectrum, it is possible to see down into each cylinder 
until the atoms become so concentrated that they virtually 
form a black wall through which vision cannot penetrate. 
This black wall in each case then becomes the " effective 
height " above the photosphere at which each spectral line 
is formed. The stronger line of the two, the one involving 
the greater number of atoms, has its effective level, or takes 
its origin, at a greater height above the photosphere than the 
weaker line of fewer atoms. The above expresses in non- 
technical language what various authorities for a number 
of years have been stating in technical terms. This simple 
scheme seems adequate to explain a great variety of solar 
phenomena. In particular, it seems easy to understand that 
within a single multiplet the spectral lines take their origin 


at different effective heights, the strongest lines at the great- 
est heights and the weakest at the least heights above the 

Through a splendid series of researches carried out at 
Mt. Wilson and elsewhere, a very imposing structure has 
been built on the assumption that Fraunhofer lines do in- 
deed take their origins at different heights above the photo- 
sphere. It seems entirely unnecessary to overthrow this 
beautiful edifice by requiring that all lines in a multiplet, no 
matter how much they differ in intensities, must all be 
found at the identically same heights. It seems equally 
unnecessary to assume that all lines in a multiplet, involving 
as they do very different numbers of atoms, must exhibit 
the same difference in wave-length between sun and vacuum 
arc. In the composite multiplet, a r 'F y D (> , the intensities 
of the strongest and weakest lines of the multiplet are as 
25:4 or 6: i, the heights are approximately as 3:1 and the 
differences sun minus vacuum arc are as i \ : i. 

From his discussion, St. John finds that the strong solar 
lines show a greater red shift than do the weaker lines except 
in the case where strong and weak lines occur in the same 
multiplet, and then the red shift must perforce be the same 
for all lines no matter how much they differ in intensities! 
On the other hand, both Burns and the writer find the greater 
shift for the strong lines even though the strong and weak 
lines belong to the same multiplet. Burns attempts, without 
much success, to explain the greater red shift of the strong 
lines as due to an " intensity equation " partly due to in- 
strumental causes. Both Burns and St. John find fault with 
the heights from the flash spectrum in that these heights are 
not a constant for all of the spectral lines of a multiplet. The 
problem of the relativity shift in the solar spectrum will 
be taken up in greater detail in Chapter XXIII. 

The theory of the formation of Fraunhofer lines and the 
problem of determining the relative abundances of elements 
in the sun's atmosphere is now (1932) receiving much at- 
tention by astrophysicists. The puzzles noted by Russell on 
page 347 show that theory and observation are not yet in 
satisfactory accord. With the perfecting of the registering 


microphotometer much observational work has been done on 
the difficult task of measuring the contours of spectral lines in 
the stars and in the sun. On account of the vastly greater dis- 
persion possible with the sun, it is evident that the chief ad- 
vances must come from solar measures. The purpose of these 
measures is to find the number, N, of atoms involved, par- 
ticularly in the formation of resonance lines of different ele- 
ments. In addition to Unsold, many other investigators have 
engaged in the problem. Wooley ' employed a clever method 
of attack by measuring the widths of the terrestrial atmos- 
pheric lines in the B-band of the solar spectrum. With the 
observations made at different times during the day, the 
known altitudes gave a ready means of deriving the relative 
numbers of atoms involved. Wooley found a variation in 
the widths of lines proportional to N* instead of N*, as ex- 
pected by Unsold 's theories. He then came to the conclusion 
that the difference between theory and observation was 
caused more by Unsold's method rather than by incorrect 
calibration by Rowland. H. H. Plaskett 2 measured the mag- 
nesium 6-triplet in the sun's spectrum. He discussed his 
measures by means of the following assumptions: (i) pure 
absorption, (2) pure scattering, and (3) combined absorp- 
tion and scattering. The third assumption seems best to fit 
both theory and observation. Pannekoek and Minnaert, 
together and separately, have made valuable contributions 
to both the theoretical and observational sides. As pointed 
out in earlier chapters, Milne has done very valuable work 
on the theoretical side. It is evident that enormous progress 
has already been made as a result of the new quantum theory, 
but the goal still seems far distant. 

In spite of the imperfect agreement between observation 
and theory, it will be very important to follow Russell's 
methods and ascertain the abundances of elements in the 
sun's atmosphere from the flash spectrum. The best pho- 
tographs for this purpose are unquestionably those of the 
Lick Observatory discussed by Menzel and the concave 
grating spectra taken by Mitchell. The information from 

1 Astro physical Journal, 73, 185, 194, 1931. 

2 Monthly Notices, R. A. S., 91, 870, 1931. 


the latter should be the more reliable, for the reasons that 
the grating spectra were photographed almost completely by 
one spectrograph with constant dispersion, while the Lick 
spectra were taken with several spectrographs and varying 
dispersions, with both fixed and moving plates; and in addi- 
tion they covered only about one-half of the region of wave- 
lengths discussed by Mitchell. 

Both Menzel and Mitchell attempted to estimate their 
intensities on the Rowland scale. In general, each agreed 
with Rowland and with each other. However, in the fore- 
going chapters the large systematic differences in intensities 
between the chromospheric intensities and those of Row- 
land have been emphasized. These differences may be 
briefly summarized as follows: (i) The hydrogen and he- 
lium lines are much stronger in the chromosphere. (2) The 
enhanced or ionized lines are likewise much stronger. The 
great increase in strength has been explained by the fact 
that the intensified lines are high-level lines. (3) With the 
neutral lines in the spectra of the metals, similar systematic 
differences between the chromosphere and Rowland are 
noted. It was shown that for any element, the strongest lines, 
or those of low excitation potential, are stronger in the chro- 
mosphere than in Rowland; those of medium excitation po- 
tential have about equal intensities in the two spectra, while 
the lines of highest excitation potential are weaker in the 

The reasons for these systematic differences for the neu- 
tral lines find a ready explanation; namely, lines of low ex- 
citation potential involve more atoms than those of the same 
element of high excitation potential. The greater number 
of atoms involved permits these lines to be detected to 
greater heights in the flash spectrum. Apparently, therefore, 
all high-level lines whether from the neutral or the ionized 
atom are stronger in the flash spectrum than in Rowland, 
while on the contrary, those of low levels are weaker than 
in Rowland. 

Remembering, therefore, that Russell's determination of 
the abundances in the sun's atmosphere depends primarily 
on the ability to calibrate the Rowland scale, it might be 


thought that on account of the great differences just noted 
between the Rowland and the chromospheric intensities, 
there might be systematic differences in the amounts of the 
elements found from the Fraunhofer spectrum and from the 
flash spectrum. In particular it might be thought that H 
and He, on account of their great strengths, would be rela- 
tively much more abundant from the chromospheric investi- 
gations. Manifestly, it is necessary also to calibrate the 
estimated intensities in the flash spectrum. The reliability 
of the final results will depend mainly on how accurately 
this calibration can be carried out. 

As Russell has pointed out, the information for each of 
the elements found in the sun depends chiefly on the strong 
lines, those of low excitation potential. In spite of the much 
greater dispersion in the Rowland spectrum than is avail- 
able in the flash spectrum, the latter has almost as complete 
information regarding the strong lines. Hence the results for 
the chromosphere from eclipse spectra should have almost 
the same accuracy as those of Russell. 

Two separate determinations have been made: one from 
the Lick photographs by Menzel, the other l by Mitchell and 
Miss Williams. The results show that within the limit of 
errors there appears to be little difference between the rela- 
tive abundances derived from the Fraunhofer spectrum and 
those obtained from the flash spectrum. In other words, the 
two determinations from the chromosphere differ from each 
other about as much as either differs from Russell's deter- 
mination. This is true for all of the elements investigated 
with the exception of H and He. For obvious reasons, the 
information from the chromospheric spectra for these two 
elements should be more reliable than from the Fraunhofer 

With the results now available, it seems probable that 
hydrogen is actually more abundant in the chromosphere, 
but whether the factor is ten, or ten thousand times, we are 
not now sure. In fact, the lightest elements, hydrogen and 
helium, still remain puzzles. In spite of the very great num- 
ber of attempts that have been made to predict from theory 

1 Publ. Leander McCormick Obsy., 5, Pt. 6, 


the intensities in the hydrogen spectrum, no very great 
success has yet been attained. 

Additional information obtained from eclipse spectra is 
the distribution of atoms at different heights, or in other 
words, the density gradients in the chromosphere. Panne- 
koek and Minnaert, from their 1927 spectra, found density 
gradients up to 2000 km for the emitting atoms of H and He. 
Their measures showed that the density of hydrogen de- 
creased much less rapidly than would be expected in an iso- 
thermal atmosphere under equilibrium conditions. The 
writer and Miss Williams (loc. tit.} have derived density 
gradients for the emitting atoms of neutral Fe and for the 
ionized elements Ti, Fe and Cr between the heights of 300 
and 2500 km. Menzel (loc. tit.) has derived a density law 
up to 2500 km by combining the measures from different 

To derive the density gradients, it is necessary to know 
the heights as accurately as possible. It is also necessary 
that the atomic origin of the lines be known and that the 
lines arise from multiplets in which the relative intensities 
of the lines can be predicted by the theoretical intensity for- 
mulae. The inter-system lines are of no use for the purpose 
for the reason that no intensity formulae are known for 
them. Further, the lines of the Balmer series of hydrogen 
cannot be used because no two lines belong to the same multi- 
plet. In the case of an element like Fe, there is a great 
amount of material available. Every multiplet in which two 
or more lines appear unblended in the flash spectrum fur- 
nishes a determination of the intensity gradient at different 
heights above the photosphere. 

Earlier in this chapter, the multiplet a 5 D y r 'D has been 
considered. The intensity formulae show that the relative 
numbers of atoms involved per unit volume are nearly pro- 
portional to the squares of the intensities. For purposes 
of illustration, we shall group together lines of equal 
heights in the multiplet and round off the predicted number 
of atoms. From the 12 lines of the multiplet we have the 









Relative number of atoms 







Height in kilometers. . . 



1 200 




With photographs of the flash spectrum taken without slit, 
measures of the chords to the tips of the cusps give the 
heights. As explained in the foregoing, a detectable blacken- 
ing of the photographic plate occurs when there is a given 
number of emitting atoms along the line of sight, a state- 
ment which is true only within a limited region of wave- 
lengths. Hence the number of atoms in line b, above the 
i50o-km level from the photosphere, is exactly equal to the 
number in line d above the looo-km level, or in line / above 
the 6oo-km level. Evidently in this multiplet, i^ times as 
many atoms are detected by the photographic plate in the 
act of emitting radiation at the i20o-km level as at the 1500- 
km level. The number of atoms per unit volume being i^ 
times, the density of concentration is therefore 50 per cent 
greater. Compared with the density at the isoo-km level, 
the density at the i2oo-km level is i^ times, at the looo-km 
level it is 3 times, at Soo-km it is 5 times, and at 6oo-km the 
density is 60 times. Hence, it is evident that from this one 
multiplet, the densities of distribution of neutral Fe atoms can 
be obtained at different heights above the photosphere. 
Other multiplets give similar information, and hence the 
density gradients may be ascertained by taking the averages 
for all the multiplets employed. Averages being taken for 
different heights, it is manifest that the most reliable results 
are derived from those spectra for which the heights are 
most accurately known. For any element like Fe where 
many multiplets are available, the material may be divided 
into different groups of low, medium and high excitation 
potential. When the number of multiplets are numerous 
enough, this same process may be carried out with different 
neutral and ionized elements. When due account has been 
taken of these matters, it is possible to derive an empirical 
density law for a number of different elements. 


In MenzePs discussion, all elements were grouped together 
while Mitchell and Miss Williams discussed the elements 
separately. The results from the two discussions agree, and 
they show that the density gradients for the various elements 
are not very different. The density gradients for neutral 
Fe and for ionized Ti are not very much greater than for 
hydrogen; and all values are much less than would be ex- 
pected in an isothermal atmosphere. 

The difference found in the distribution of the atoms of Fe 
of low excitation potential, as compared with that of the 
atoms of higher excitation potential, might be ascribed to 
a large temperature gradient in the chromosphere were it 
not for the fact that the Ti-\- data shows no such effect. 
The same effect is undoubtedly present in the Fc data at 
much lower levels in the regions between the effective 
levels of the reversing layer and the chromosphere. Menzel 
ascribes this to a temperature gradient, but in view of the 
fact that the Ti + data shows no such effect in this region, 
even as it showed no effect at greater heights, it is unsatisfac- 
tory to assume an explanation which brings the Fe data into 
step but at the same time throws the Ti + data out of step. 

But why are the density gradients so much smaller than 
expected? What is the supporting mechanism? Many theo- 
ries have been proposed. These have been examined in de- 
tail by Menzel. Here it will be possible to touch on them 
very briefly, on account of limitations of space and also for 
the reason that none of the theories satisfactorily explain the 
observed facts. Milne (loc. cit.) has a beautiful theory in 
regard to ionized calcium. The atoms fall toward the sun 
under the influence of gravity until a quantum of energy of 
the frequency of the H or K line is absorbed. The momen- 
tum associated with the light-quantum tosses the atom high 
into the chromosphere again, after which once again it 
begins to fall. This process is repeated indefinitely. This 
theory may be applied to the other alkaline earths, Mg, Sr 
and Ba, but it is not applicable to any of the other elements ; 
some other force must support them. As Rosseland ! was 
the first to suggest, there is a great turbulence in the sun's 

1 Monthly Notices, R. A. S., 88, 377, 1928. 


atmosphere in virtue of which the heavy elements will be 
tossed to great heights in the chromosphere along with the 
light elements. This would account for the fact that the 
relative abundances found in the reversing layer, that is, in 
the lower and denser layers of the chromosphere, are prac- 
tically the same as found in the higher reaches of the chro- 
mosphere from eclipse spectra. Turbulence would also 
account for the similar density gradients of dissimilar 

Gurney's theory 1 is most interesting, especially as it may 
give a solution for the helium puzzle. Since there are so 
many Ca -f- atoms in the high chromosphere, a great number 
of them will become doubly ionized. But the doubly ionized 
Ca ion is incapable of supporting itself by absorbing light- 
quanta, for the reason that its ultimate lines lie in a spectral 
region where the number of light-quanta emitted by the sun 
are relatively few. Hence the Ca + + ion will drop back to 
the photosphere and in so doing it will attain a tremendous 
kinetic energy, enough to excite the visible helium lines if it 
should collide with a He atom, or even enough to ionize the 
He and excite the interesting line found in the flash spectrum 
at wave-length 4686. 

1 Ibid., 89, 49, 1928. 




corona still remains exclusively an eclipse phe- 
nomenon. In spite of the amazing achievements 
of modern science which at times seems to be able 
almost to accomplish the impossible, no success has attended 
the efforts made to observe the corona outside of an eclipse. 

On account of the dramatic character of the phenomenon 
and of the great interest in eclipses felt by both the astrono- 
mer and the general public, each eclipse as it comes is en- 
thusiastically observed ; but the truth must be told that suc- 
cess commensurate with the labor involved is not always 
forthcoming. " The problems of the corona are many, and 
few of them can be said to have approached solution. Im- 
portant facts concerning it have been established, but these 
facts are more or less isolated, and in general their relations 
to each other are unknown. The paucity of results obtained 
thus far is due primarily to the unique condition that the 
most assiduous of eclipse observers can scarcely hope for 
more than an hour of totality with clear skies, in his entire 
lifetime/' In this and the following chapter we shall attempt 
to show how the amazing progress of recent years in every 
branch of solar research has affected the solution of coronal 

Concerning the total eclipse of March 29, 1652, seen in 
Ireland, Dr. Wyberd writes, " The moon suddenly threw it- 
self within the solar disk with such agility that it seemed 
to go round like an upper mill-stone. The sun then appeared 
around the limb, affording a pleasant and remarkable spec- 
tacle of rotation." There seemed to be a widespread notion 
among the early observers of eclipses that during totality 
very rapid motions took place within the corona and 
that the corona was some sort of a modern fireworks 



display with brilliant scintillations and sudden changes. In 
fact, the early notion seems not to have entirely vanished 
in the enlightened and scientific age of the twentieth 

No eclipse expedition worthy of the name will be fully 
equipped unless it has as part of its program the securing 
of large scale photographs of the corona. Astronomy of the 
future needs, as it has in the past, to secure good photographs 
of every possible eclipse. The Lick Observatory has the 
most complete series in existence, beginning with the eclipse 
of 1893. The Lick photographs have always been secured by 
pointing the camera directly at the sun, the method devised 
by Schaeberle, and a uniform focal length of forty feet has 
been employed. This permanent record of the past is always 
available for purposes of comparison. 

The 5-inch aperture, 4O-foot focus represents a ratio of 
aperture to focal length of 1:96. The commercial photog- 
rapher would look aghast if he were compelled to work with 
such a slow camera. For investigating the details of the 
inner corona, a large scale image is necessary, rendered pos- 
sible by great focal length. Up to the present, the largest 
scale has been secured at the eclipse of 1900 through the use 
of a horizontal camera of 12 inches aperture and 135 feet 
focal length, a ratio a:/ =1:135. The brightness of the 
inner corona permits the employment of short exposures even 
with the small ratio of aperture to focal length necessitated 
by cameras of great focal length. 

Two methods of mounting cameras of great length are 
available, either that of pointing the objective directly at the 
sun, or using the horizontal telescope with coelostat mirror. 
The former has distinct advantages over the latter, but the 
erection becomes increasingly difficult as focal lengths are 
augmented. Miller of Swarthmore College has been very 
successful with focal lengths of 63 feet. For such a mount- 
ing, a double tower is necessary, an inner one to carry the 
objective, and an outer tower to protect the whole from jars 
caused by the wind. With this type of mounting the plate is 
moved by a clock to counteract the diurnal motion. With 
the Swarthmore tower-telescope on Niuafoou Island, at the 


most recent eclipse, photographs with superb definition were 
secured by Marriott. 

The horizontal telescope is much easier to construct and 
erect in the field. The tube may be made of heavy paper and 
the plate holders carried by a heavy framework surrounded 
by a dark room. Care should be taken that the tube is not 
too near the ground and that the tube is protected from the 
direct rays of the sun by a canvas or paper shelter. The 
coelostat is the best form of mounting for the plane mirror. 
Compared with direct mounting, as was effectively shown at 
the eclipse of 1919, the horizontal telescope has the great 
disadvantage that the mirror is sensitive to changes of tem- 
perature and it may alter its shape and become warped on 
eclipse day when exposed to the sun's rays. Ordinarily, focus 
is obtained for eclipse cameras by star trails. If possible a 
warm night, differing as little as possible from the expected 
temperature at eclipse time, should be utilized. It is impera- 
tively necessary to keep the mirror protected from the sun's 
rays on the day of the eclipse until a very few minutes before 
totality. One should avoid the horizontal telescope if pos- 
sible, mainly on account of the uncertainties connected with 
the mirror and its lack of permanent figure. At the Niuafoou 
Island eclipse, good photographs were secured with the 65- 
foot horizontal telescope belonging to the U. S. Naval Ob- 
servatory. The lens is as good a one as the Swarthmore lens 
of nearly equal focal length used in the tower-telescope and, 
moreover, Marriott cared for the details of both instruments 
and the development of the plates. The definition of the 
tower plates, however, was better than those taken with 
the horizontal telescope. For the purpose of portraying the 
structural details of the corona it is evident that the greater 
the focal length utilized the more valuable will be the results. 
Cameras of medium and small size have not been superseded 
by the large instruments and still have useful functions at 
eclipses. Larger ratios of aperture to focal length are possi- 
ble than with those of the largest size. Such cameras are 
more rapid and, consequently, are useful in securing the faint 
extensions of the outlying corona. They are especially valu- 
able in photometric observations which will be taken up in 


detail below. Every precaution should be taken to reduce 
to a minimum the effects of halation and reflections from the 
glass-side of the plates. All plates used for photographing 
the corona should be " backed " with some absorbing material 
on the glass side. Some eclipse observers have had excellent 
success from using double-coated or triple-coated plates. 

In attempting to photograph the corona on small or large 
scale, it goes without saying that the greatest care should 
be exercised to secure the most perfect definition, and also 
that the photographic manipulation should be conducted 
so as to bring out of the plates as much of the wealth of 
detail as possible. Development may be carried out in such 
a way as to attain very different photographic effects. If 
contrast is needed, as in spectrum work, a " hard " develop- 
ment is required. There are many developers suited for this 
sort of work with which anyone doing spectrum work is en- 
tirely familiar. For developing the corona, however, where 
it is necessary to bring out as many of the fine details as pos- 
sible, and where an attempt should be made to minimize the 
contrast between the bright inner corona and the faint outer 
corona, an entirely different kind of developer and develop- 
ment is necessary. Old-fashioned " pyro " is probably the 
best form of developer to use for this purpose, and one would 
do best to start with a very weak solution and proceed gradu- 
ally. The proper development of each plate will take at least 
an hour. Undoubtedly many well exposed coronal photo- 
graphs have been spoiled through improper care in develop- 
ment. The technique of development of photographs is so 
well known that no attempt at further details will be given 

In reviewing the scientific work accomplished at solar 
eclipses, one is immediately struck by the fact that success 
has been greater in the direct photography of the corona than 
in any other branch of eclipse investigations. The self-evident 
reason is that every well-equipped eclipse expedition, no mat- 
ter what their other program, attempts to photograph the 
corona. However, there is another reason not so apparent, 
which is that perfection of focus and seeing, although de- 
sirable, are not absolutely essential in obtaining successful 


photographs of coronal details. A lack of perfect focus plays 
havoc with spectroscopic photographs but detracts little from 
the corona for the reason that its structure is nebulous or 
filmy in detail. These remarks should not be interpreted 
to mean that one should be satisfied with anything short of 
absolute perfection in the determination of the best focus. 

The general form of the corona can be predicted in advance 
of the eclipse. At sun-spot minimum are found the long 
equatorial streamers and the short plume-like polar brushes 
which were well seen at the eclipse of 1900 (p. 164), or in 
1922 (p. 365). At sun-spot maximum the corona is nearly 
circular in shape, thus resembling a gigantic dahlia. How 
closely are these typical shapes connected with the sun-spot 
curve? It is rather curious to find that in the thirty years 
following 1893, when large-scale photographs of the corona 
were first successful, sixteen total eclipses were observed 
either near maximum or minimum of spots, or were distrib- 
uted along the branch of the curve descending from maxi- 
mum. The eclipse of 1914 was the only corona seen on the 
ascending branch of the sun-spot curve, the corona of 1916 
having been obscured by clouds. Fortunately for our knowl- 
edge of this subject, the eclipses of 1925, 1926 and 1927 were 
all on the ascending part of the curve. With the eclipses of 
1929 and 1930 on the descending branch of the sun-spot curve 
near maximum, our information regarding the shape of the 
corona can now be subjected to more exact scrutiny. We ask 
ourselves the question whether the typical coronal shapes 
take place exactly at maximum and minimum, or whether 
they start before or after the times of maximum and mini- 
mum. Take one particular eclipse. Sun-spot maximum oc- 
curred in August, 1917. The eclipse of June, 1918, did not 
show the typical corona of sun-spot maximum, for the polar 
streamers were shorter than those near the equator, and the 
corona departed more from the circular shape than was ac- 
tually anticipated. Generally when the maximum of spots 
is past, the streamers draw away from the poles, and the long- 
est rays are found in the sun-spot zones, making the corona 
rectangular in appearance. . 

Although it may be said with truth that the shapes of the 







corona at minimum of sun-spots all resemble each other in 
having long equatorial streamers and pronounced polar 
brushes, yet each corona has its own particular features, its 
own peculiar structure. The coronas of 1878, 1889, I 9 
and 1922, all taken at minimum phase of sun-spots, could 
never be mistaken one for the other. In fact there were two 
eclipses in the year 1889, on January i and December 22, 
with very pronounced alterations in shape in the coronas. 
These variations in form, as pointed out by Wesley, 1 were 
precisely in accordance with the change to be expected with 
the sun returning toward a condition of spot activity. The 
shapes of the coronas from 1922 to 1927, inclusive, furnish 
interesting comparisons. The eclipse of September 14, 1922 
represented the typical minimum corona. The eclipse of 
September 10, 1923 exhibited the same features. Even on 
January 24, 1925, there was still the typical minimum shape, 
of pronounced polar brushes, but the equatorial streamers 
were not so long as those of 1922. To the right of the verti- 
cal, however, there was a long pointed shaft of light giving 
mute evidence that the time of minimum was well past. On 
January 14, 1926, the corona had lost all of the characteris- 
tics of the minimum type and closely resembled that associ- 
ated with maximum of spots, though the time was two and 
one half years before the expected maximum of spots. The 
corona of June 29, 1927 was distinctly that of maximum 

Corona times near sun-spot maxima resemble each other 
even less than do the different aureoles at sun-spot minima. 
The coronas corresponding to maximum of spots are, it is 
true, approximately circular in outline, but prominent rays 
and streamers shoot out at various angles. We are forced 
therefore to the conclusion that changes are continuously go- 
ing on in the corona. How rapid are these changes? Can they 
be detected from photographs taken with the same camera, 
at the beginning and ending of a total eclipse? It is possible 
that the few short minutes of totality may afford too short an 
interval to permit these changes to be detected with cer- 
tainty, from photographs taken at any one location, but mo- 

1 Observatory, 13, 105, 1890. 


tions might possibly be detected from photographs secured 
at widely separated stations at the same eclipse. 

Any theories dealing with coronal structures must depend 
on the measured changes within the corona itself , and it is 
therefore imperatively necessary that reliable information 
be secured regarding these motions. The importance of 
this problem has been fully recognized by eclipse observers 
for many a long year. Manifestly, no real progress was 
possible as long as it was necessary to depend for informa- 
tion on drawings of the corona. In fact, sketches made dur- 
ing the time of an eclipse have been a continued disappoint- 
ment, but the non-success was no greater than what should 
have been expected. Even a skillful draughtsman subjected 
to the excitement and unfamiliarity of a total eclipse could 
not be expected to see and draw, in the hurried interval of 
a couple of minutes, more than a few details of coronal struc- 
ture. These details of necessity must be exaggerated in the 
drawing. A sketch made by a second person, as well trained 
as the first draughtsman and secured at the same time and 
location, probably would stress other details regarded as 
important. But to make matters worse, most of the draw- 
ings of the corona made in the past were secured by men 
who were observing their first eclipse, and frequently these 
very men had little experience or skill in the making of 
sketches. With the advent of the photographic plate, the 
wholesale drawing of the corona during totality has been 
pushed more and more into the background. 

To detect motions in the corona with the greatest cer- 
tainty, several prerequisites are necessary: First, the photo- 
graphs to be measured should be on as large a scale as pos- 
sible; second, the interval in time should be as great as 
possible; and third, the photographs to be compared should 
be secured with cameras of nearly the same focal length and 
they should resemble each other in general appearance as 
much as possible. Plates developed with different effects of 
contrast and taken under different conditions of seeing cannot 
furnish motions with the highest degree of precision. 

Attempts were made as early as 1889 to secure the neces- 
sary information by the comparison of photographs at the 


eclipse of December 22. An interval of two and a half hours 
elapsed between the time of totality at Cayenne and Cape 
Ledo in West Africa. Unfortunately, clouds were experi- 
enced at the latter station. As a result of the long duration 
of totality at the eclipse of 1901, an exceptionally favorable 
opportunity existed for determining motions from observa- 
tions at a single station. Perrine 1 secured successful photo- 
graphs with the 40-foot camera. Measures of short exposure 
negatives, taken near the beginning and end of totality, 
showed no displacements of coronal masses in the interval 
of a little more than five minutes. On account of the accu- 
racy of the measurements, it was possible to have measured 
with certainty a velocity of 20 miles per second across the line 
of sight. Motions should have been suspected if they had 
been as great as 12 or 15 miles per second. The 1901 pho- 
tographs were near sun-spot minimum. At the eclipse of 
1905, Hansky secured photographs on the same scale as Per- 
rine's photographs/ The corona was one of typical maxi- 
mum form and presented few conditions for the best results. 
The corona was very intense, especially near the solar sur- 
face, and its rays, being projected from all sides of the disk 
of the sun, so superposed themselves the one upon the other 
and became so entangled that it was almost impossible to 
distinguish the same details on successive photographs. By 
taking the original negatives and making glass positives, by 
proper shading and by local development, much detail was 
secured in the corona. From the measures of 43 separate 
rays, information was obtained concerning the motions of 
the coronal material within the period of totality lasting a 
little more than three minutes. The velocities determined 
were little greater than the errors of observation. None of 
the velocities investigated exceeded 25 km per second. 

In accordance with the plans of the Lick Observatory of 
always attacking eclipse problems of the greatest importance, 
an attempt was made at the eclipse of 1905 to detect changes 
in the coronal structure by establishing three stations at 
widely separated localities and securing large-scale photo- 
graphs at each station. Parties were sent to Labrador, to 

1 Lick Observatory Bulletin, i, 151, 1902. 2 Mitt. Pulk., 2, No. 19, 107. 


Spain and to Egypt. Unfortunately for the success of the 
scheme, cloudy weather prevailed in Labrador where were 
stationed Curtis and Stebbins. Thin clouds were met in 
Spain, but they did not greatly interfere with the photo- 
graphs secured by Campbell and Perrine. Clear skies greeted 
Hussey in Egypt where totality occurred 70 minutes later 
than in Spain. A careful comparison was made of the photo- 
graphs 1 secured at the two stations. A number of fairly 
well-defined nuclei were found, both east and west of the 
sun. Details of structure within the nuclei appeared to 
change, but the nuclei as a whole remained in the same posi- 
tion. Measures of the greatest accuracy were impossible 
on account of the poor definition of the Egyptian plates 
caused by poor seeing. The conclusion was that " the masses 
in question could not have moved so much as one mile per 
second during the interval of 4200 seconds. Greater speeds 
might well have occurred within the principal coronal stream- 
ers, or within some of the arched forms which enclose promi- 
nences, without our having detected them; for their structure 
is quite uniform, and well-defined nuclei are absent/ 7 

The American eclipse of 1918 afforded an additional op- 
portunity by comparing plates taken at three separate sta- 
tions where were located expeditions from Lick, Lowell and 
Sproul Observatories. At Goldendale, Washington, the focal 
length was 40 feet, at Syracuse, Kansas, the camera was 38.7 
feet in length, while at Brandon, Colorado, the camera was 
63 feet long. One plate secured by each of the first two 
cameras and two plates with the 63-foot instrument were 
compared. Each of the three cameras pointed directly at the 
sun. There were no definite nuclei or other distinctive fea- 
tures present in the streamers that were sufficiently well de- 
fined to be used as points of measurement by Miller." There 
were, however, three arches surrounding three prominences 
and attention was confined to these. The first arch was 
around the " Pyramid " or " Eagle " prominence in the north- 
eastern quadrant. Towards the pole side of this prominence 
there were four well-defined arches. The second arch was 
around a prominence in the SE quadrant, while the third 

1 Lick Observatory Bulletin, 4, 121, 1907. 2 Publ. A. S. P., 32, 207, 1920. 


arch was near the " Heliosaurus " prominence. Each of 
the three prominences displayed four separate arches, and 
the positions of one arch only of each were measured. These 
measures gave fairly accordant results and seemed to show 
that the arches had changed in the twenty-six minutes' in- 
terval between the Lick and the Sproul photographs. The 
average rate of speed at which these arches receded from the 
sun was about ten miles per second. 

At the eclipse of 1926, Horn-D'Arturo compared photo- 
graphs taken by the Italian expedition in East Africa with 
those taken by the British expedition in Sumatra two and a 
half hours later. He measured the coronal domes at position 
angle 335 and found that if they moved at all, their motions 
were not greater than 2 km per second. Von Kliiber l reached 
a similar conclusion from measures on the photographs of the 
German expedition in Sumatra compared with Horn-D'Ar- 
turo's measures. 

Horn-D'Arturo found that on the photographs taken in 
Africa there were well-defined streamers at position angles 
350 and 320, both traceable to heights greater than one 
solar radius, whereas on the Sumatra photographs the 
streamer at 350 had disappeared, the streamer at 320 was 
intensified and a new, well-defined, very straight streamer 
had appeared at 325, traceable to a height of ii solar radii. 
If the new streamer be interpreted as matter just ejected 
from the sun, then the matter must have been moving at 
least 50 km per second. 

From the photographs compared by Miller, it was evident 
that the polar rays east of the axis of symmetry are curved 
away much less than the rays to the west, the difference 
not being due to the effect of projection. It is therefore 
evident that the velocities with which material is ejected 
from the sun to form the coronal streamers must be very 
small compared with the motions with which we are fa- 
miliar in the prominences. Any theory dealing with the 
formation of the coronal structure must take cognizance 
of these moderate velocities of ejection. 

The corona owes its entrancing beauty to the far-flung, 

1 Zcits. jur Astrophysik, 4, i, 1932. 


pearly-white streamers, so conspicous in each corona ob- 
served. The contrast with the rosy-red prominences close 
to the edge of the black moon makes a never-to-be-forgotten 
spectacle. In addition to the streamers, there are other 
features of the corona that deserve attention, and it is neces- 
sary to find out their relations to solar phenomena such as 
prominences, sun-spots and faculae. " Arches," " hoods " 
or " striated cones " have been observed since Cleveland 
Abbe first saw them at the eclipse of 1869. They are 
specially conspicuous at the time of sun-spot maximum. 
The first good photographs obtained of them were by 
Schaeberle at the eclipse of April, 1893. The coronal arches 
were noted by him 1 to be associated with prominences. 
Miss Clerke" draws the conclusion that " Each pearly pa- 
vilion is erected over a red flame. Coincidences of the kind 
are of perpetual occurrence." These hoods were specially 
marked in the coronas of 1896 and 1898, but were prac- 
tically missing from the minimum type of corona of the 
year 1900. 

At the eclipse of 1901, a different kind of " disturbance " 
was noted in the photographs secured by Perrine, which 
consisted of a conspicuous center, apparently at the sun's 
limb, from which strong streamers stretched out to great 
distances from the edge of the sun. This disturbance was 
all the more interesting for the reason that Perrine found 
it to take its origin over the region of a prominent sun-spot. 
In fact, this spot was the only one known to exist on the sun 
at the time. Similar disturbances were seen in the photo- 
graphs of the eclipse of 1918, but Campbell and Moore 
could find no relation between them and any sun-spot or 
chromospheric phenomena known to exist on the sun. As 
already stated, the arches were specially prominent at this 
recent eclipse and they extended to much greater distances 
from the sun's limb than they did in 1893, although both 
eclipses were near the sun-spot maximum. In 1923, Miller 
photographed a disturbed area in the corona intimately asso- 
ciated with a spot near the sun's limb, shown at page 209. As 

1 Contributions from the Lick Observatory , 4, 94, 1895. 

2 Problems in Astrophysics, 129, 1903. 


a whole, disturbances are much more conspicuous in the coro- 
nas seen at maximum of sun-spots for the reason that promi- 
nences are much more numerous and active and are found in 
all heliographic latitudes, while at sun-spot minimum the 
prominences are feeble and are confined to zones near the 
solar equator. 

Much valuable information regarding coronal disturbances 
was obtained 3 at Niuafoou Island in 1930. This was the first 
time in the history of eclipses that so much detailed structure 
was visible in the spectral lines of the corona, this structure 
being made possible by the fine definition of the concave grat- 
ing spectra taken without slit. A comparison of all the 1930 
grating spectrograms, but more particularly a study of the 
H and K lines, reveals the interesting fact that the sun, two 
years after sun-spot maximum, was in a condition of great 
activity on the day of the eclipse. The spectra and the 
direct photographs taken by the 63 -foot tower telescope by 
Marriott show prominences completely encircling the sun. 
Comparisons were also made with spectroheliograms on four 
successive days, taken at Mt. Wilson or Kodaikanal. The 
superb definition of the eclipse photographs shows both the 
prominences and the inner corona in great beauty. 

A comparison of all the photographs, the spectrohelio- 
grams with the eclipse plates both direct and spectrographic, 
shows that the sun was very active not only at eclipse time 
but throughout the whole period of nearly four days covered 
by the plates. At the time of totality there was a promi- 
nence almost exactly at the south point of the sun extending 
to a height of 25,000 km. At this time the axis of rotation 
passed 26 to the west of the south point. The photographs, 
direct and spectrographic, show prominences even at the 
north and south poles of the sun's axis. 

Marriott's photographs and the eclipse spectra exhibit 
that the whole southeast quadrant was a tremendously 
stormy region on the sun. The center of the longest stream- 
ers of the whole corona was situated 30 to the east of the 
south point of the sun, or at position angle 150. Inter- 
twining the coronal streamers is a beautiful series of coronal 

1 Publ. Leander McCormick Obsy., 5, 155; Astroph. Jour,, 75, i, 1932. 


domes. At totality, at the base of the long streamers and 
domes is an extended group of prominences centered at 150 
of position angle. These prominences, though apparently 
very active as shown by the eclipse photographs, did not 
reach the greatest heights on the sun. These greatest heights 
were actually attained by prominences also in the southeast 
quadrant but at 115 of position angle. This is the location 
of the conspicious feature called the " strawberry dome," with 
its magnificent series of arches and with its delicate inter- 
twined structure of filamentous details. Facing page 373, at 
A and B are found drawings from the eastern edge of the 
sun from the region where the " strawberry-shaped " dis- 
turbance is found on Marriott's photographs. At A are 
drawings from the K and Ha lines of the spectra of the 
chromosphere, and at B is the same region from the direct 
photographs and from the lines 5303 and 6374 of coronium. 
At C on the western edge of the sun, the detail in the 5303 
line resembled the cocoanut trees which grew in great pro- 
fusion on " Tin-Can " Island. 

For many long years all eclipse observers have called at- 
tention to the connection between coronal streamers and 
prominences. This dependence of streamers and prominence 
activity is abundantly verified in 1930, but this eclipse dem- 
onstrates the fact that the longest coronal streamers, on 
which the shape of the corona more or less depends are not 
necessarily connected with the prominences which at the 
time of the eclipse are of greatest height. 

The long streamers of the corona are best seen near the 
time of minimum of sun-spots. The splendid series of photo- 
graphs secured by Lick Observatory expeditions have been 
utilized by Moore and Baker to measure the direction of the 
axis of symmetry of the polar brushes, photographs at five 
separate eclipses exhibiting the polar rays sufficiently well- 
determined for the purpose. The direction of the sun's 
equator is inclined about 7 to the ecliptic and 26 16' to 
the terrestrial equator. The measures given in the following 
table show that the axis of symmetry of the polar rays coin- 
cides, within errors of measurement, with the rotation axis 
of the sun. 




C lH 

O O 

W M 




A and B refer to the eastern section of the sun shown in Plates I and II, 
and C to the western section. (Please note that all positions in C are inverted.) 

In A are given the K line of Ca-f- and the Ha line of the chromosphere 
from spectra taken immediately following the first flash. In B and C are 
given (from left to right) details from Marriott's photographs near the 
beginning (B) and end (C) of totality, and the structure in the coronal 
rings of spectra at X 5303 and X 6374, respectively. 





(Position angles measured from North point of Sun) 

Date of Eclipse 

Axis of Polar Streamers 









1898 Jan. 21 

9.4 W 

9-3 W 

9.4 W 

8.2 W 

1.2 W 

IQOO May 28 

17.1 W 

17.6 W 

17.4 W 

16.9 W 

0.5 W 

1901 May 18 

21. 5 W 

20.7 W 

21. 1 W 

20.3 W 

0.8 W 

1908 Jan. 3 

0.9 W 

0.8 E 


1.2 E 

1.2 W 

1922 Sept. 21 

24.4 E 

24.8 E 

24.6 E 

25.0 E 

0.4 W 


0.8 W 

Facing page 373, line i following caption, for in Plates I 
mcl IF read facing pages 8 and 13 

granted that every spectral line seen at mid-totality, when 
to the eye there was no visible trace of the chromosphere, 
must perforce take its origin in the corona. It was with a 
great shock of surprise that H and K of calcium were seen in 
1882 projected on the black face of the moon, where presum- 
ably there is no light at all. About thirty bright lines were 
photographed by Schuster and by Abney at the eclipses of 
1882, 1883 and 1886. Even as late as 1893, the H and K 
lines were assumed to belong to the corona. 

Knowledge of the spectrum of the corona may almost be 
said to begin with the eclipse of 1893, when Fowler with a 
prismatic camera photographed nine rings, all of which agree 
in position with lines reported for the 1886 eclipse. At the 
same eclipse, Deslandres photographed three coronal lines 
at 3987, 4086 and 4231 A, and also three others at 3164, 
3170 and 3237 A, in the ultra-violet. These last three have 
not been observed since by others, nor have two of Fowler's 
nine. It was not until 1898 that coronal spectra had acquired 
sufficient precision to distinguish between the chromospheric 


line 1474 K at wave-length 5317 and the " coronium " line 
fourteen angstroms farther to the violet, at 5303. Photo- 
graphs were secured at the eclipse by Fowler, by Campbell, 
by Naegamwala, and by Newall and Hills. Additional pho- 
tographs were secured by Frost in 1900, by Dyson in 1900, 
1901 and 1905, by the U. S. Naval Observatory parties in 
1900, 1901 and 1905, by Fowler and Lockyer in 1900, and by 
Lockyer and by Campbell in 1905. More recently the Lick 
expeditions have secured good photographs, by Lewis, and 
by Campbell and Albrecht in 1908, and by Lewis, by Camp- 
bell and Moore in 1918 and again in 1922. More recently 
excellent spectra have been secured, in 1925 by Curtis and 
Burns, in 1926 by Davidson and Stratton, in 1929 by Gro- 
trian, and in 1930 by Mitchell. 

At the eclipse of 1918, Slipher found on his photographs 
the green coronium line extending across the space occupied 
by the moon's image due to scattering of light in the 
earth's atmosphere. At the eclipse of August, 1914, Car- 
rasco, and Bosler and Block, independently discovered a new 
coronal line in the red, at wave-length 6374 A. In Lick Ob- 
servatory Bulletin, 10, i, 1918, Campbell and Mopre assem- 
ble all the reliable observations of coronal wave-lengths de- 
termined since the eclipse of 1893. These values for forty 
lines are given in the second edition of this book on page 329. 
As an illustration of the precision of knowledge at that time 
regarding coronal wave-lengths may be cited the measures 
for the strongest line, that in the green. Before the recent 
eclipse, the most accurate results came from the 1918 obser- 
vations, by Campbell and Moore, and by Adams, St. John and 
Miss Ware. Uncorrected for rotation (0.035 A), and cor- 
rected (assuming the corona rotates at the same rate as the 
photosphere ) , the values in international angstroms are given : 

Uncorrected Corrected 

Campbell and Moore 5302.83 5302.80 

Adams, St. John and Ware 5303.02 5303.06 

The difference in the corrected values is 0.26 A, while 
the mean of the two measures is 5302.93. The mean value is 
in excellent agreement with the 1930 result of 5302.91 A 
from the measures by Mitchell of eight spectra. Six spectra 


at the 1930 eclipse give 6374.28 A, for the line discovered 
at the 1914 eclipse. 

It might be said that the corona exhibits three separate 
spectra: first, the bright-line spectrum of " coronium " exist- 
ing only in the inner corona and extending on the average to 
about 5' or 200,000 km from the sun's edge; second, the con- 
tinuous spectrum of the middle corona; and, third, the Fraun- 
hofer lines showing feebly in the outer corona. 

In the Revised Rowland Tables, page 226, 1928, is given 
the list of bright lines attributed to the corona by Campbell 
and Moore but revised to include 1 the 1926 eclipse. In 
Zeitschrijt fur Astrophysik, 2, 106, 1931, Grotrian gives the 
results of the Potsdam expedition to Sumatra of May 9, 1929. 
In the following table, the 1930 eclipse is also included. 
After eliminating the lines due to the high chromosphere, 
there are only eighteen lines out of forty which are now 
believed to be truly coronal in origin, although a line or two 
may still be suspicious. The first column gives the wave- 
lengths in international units. Those at the top of the table, 
given to two places of decimals, are from the 1926 eclipse 
by Davidson and Stratton. The values of 5302.91 and 
6374.28 are from the 1930 spectra. The line at 6704 was 
discovered at the 1929 eclipse and verified by the 1926 eclipse 
photographs. The line at 6776 announced by Mitchell at the 
1930 eclipse needs to be verified. The intensities given in 
the remaining columns are estimates where the strongest 
lines, 3388 in the violet and 5303 in the green, are usually 
assumed to be of intensity 20. The intensities given in the 
table are those assigned by the observers except for the 1918 
eclipse where Moore and Campbell's estimates are each mul- 
tiplied by the factor 2 so that the line 5303 may have an 
intensity 20. For the 1922 eclipse, the intensities have been 
furnished by letter through the kindness of Moore. In the 
second column an average intensity is assigned from the mean 
of all the eclipses. 

The intensities of the coronal lines in the table must be 
considered to be on a relative scale and not absolute for the 

1 Davidson and Stratton, Memoirs of the Royal Astronomical Society, 64, 105, 



reason that 5303 varies from eclipse to eclipse and is not con- 
stant as was assumed in forming the tabular values. 

Before taking up the striking differences in the intensities 
from eclipse to eclipse shown by the table, one must consider 
the factors upon which the coronal intensities depend. They 
may be grouped under four heads: (i) color sensitivity of 
the photographic plate, (2) the dispersion, (3) prism or grat- 
ing, and (4) slit or slitless. The improvements in plates 
in recent years, making them more sensitive at the red end, 
are mainly responsible for the discovery of the lines in the 
red. Prisms of glass absorb at the violet end of the spectrum ; 


I A. 



ton, and 


1901 - 










28 . 



^87 96.. 









3454 U 


















^642 87.. 









3800 77 












3986 88 . 













4086 2Q . 






















43 1 1 


































^02 91 . 

















6m 28 









gratings (of speculum metal) and the silver-on-glass re- 
flectors absorb in the ultra-violet. Increase of dispersion 
spreads the monochromatic images farther apart and weakens 
the continuous spectrum, with the result that the emission 
lines by contrast are easier to see on the photographs taken 
with higher dispersion. 

At the 1930 eclipse on Niuafoou there was a striking dif- 
ference between the coronal spectra taken with the concave 
gratings and with the prismatic camera of Dr. C. E. Adams 
of the New Zealand expedition. The prismatic camera gave 
a very brilliant spectrum in fact, more brilliant than that 
of either concave grating. Both forms of spectrographs were 


without slits. The dispersion of the gratings, however, was 
ten times the average dispersion of the prismatic spectrum. 
Many coronal rings were seen on the grating photographs, 
the lines 5303 and 6374 being well visible even on the brief 
exposure for the second flash. Oft the prismatic spectra, on 
the contrary, the only coronal line visible on any of the photo- 
graphs was 5303, and even this line was seen with great 
difficulty. It is readily seen therefore that the dispersion 
employed has a very important bearing on the intensities, and 
hence any conclusions drawn from variations in intensities 
between the lines of one eclipse and another have very little 
meaning unless the instrumental conditions are fairly con- 

The specially interesting features of the coronal spectra 
of 1930 are the large dispersion and absence of slit. On the 
photographs, the image of the sun is 14.5 mm in diameter and 
the dispersion 10.9 angstroms per mm. All of the coronal 
rings show much interesting detailed structure, but this is 
specially true for the strong 5303 and 6374 rings. On the cuts, 
facing page 372, the vertical line coincides approximately 
with the axis of rotation of the sun. One can see at a glance 
that if a slit had been used in 1930 passing radially through 
the center of the sun, the intensity of 5303 (or 6374) would 
have depended on the position angle of the slit. If placed 
near the solar axis, the coronal lines would have been feeble 
while if it had happened to pass over a more brilliant portion 
(which could not have been told in advance) the intensity 
would have been much greater. Hence, for discussing the 
distribution of coronium about the sun, slitless spectra (if 
obtained with large dispersion and good definition) are vastly 
superior to spectra taken with a slit. 

With these instrumental sources of variations in intensities 
clearly in mind, let us look at the estimates of intensities in 
the foregoing table. There are several striking peculiarities. 
Many of these, and others, have already been noted by Camp- 
bell and Moore. The most remarkable involves the line at 
3601 A. It was not observed by any one until the eclipse 
of 1908, in spite of the fact that it is one of the strongest lines 
of the coronal spectrum and the quartz spectrographs were 


adequate to secure it. The photographs by Mitchell at the 
three eclipses of 1905, 1925 and 1930 were all taken with the 
same spectrograph. The plates of the three eclipses have 
about the same sensitivity in the violet. No trace of 3601 is 
found in 1905 although 3388 and 3454 farther to the violet 
are strong lines. By way of contrast, in 1925 and 1930 the 
line 3601 was very strong. (Other peculiarities of this line 
will be referred to later.) Likewise, 4086 was not observed 
before 1908. At the 1929 eclipse Grotrian reports three lines 
missing, 3643, 4586 and 5536. The last line of the three 
is the weakest of the whole coronium spectrum; it is, more- 
over, near a Fe + line in the chromosphere of 5534.85 and 
intensity 15 in the flash spectrum. Davidson and Stratton 
ascribe this line to the corona. The first of the three lines at 
3643 is on the average a fairly strong line. Mitchell could 
not be sure of it from his 1925 spectra, however, though it 
was well visible in 1905 and 1930. 

One's first thought presented for an explanation of these 
peculiarities is that the intensities vary, and, having settled 
this to one's satisfaction, the next thing is to try and con- 
nect the variation with the sun-spot cycle. There we strike 
a snag. The years 1900 and 1901 had coronas of minimum 
type while the 1905 eclipse was of maximum type, and yet 
3601 was invisible in all three years and, moreover, it has 
been a strong line in every eclipse without exception since 

The first step towards finding the origin of the bright 
lines in the corona evidently will be to explain the peculiari- 
ties of intensities just noted by finding relationships between 
the different lines. To this end, Stratton and Davidson ' 
summarize the similarities of coronal structure noted pre- 
viously by others. Lockyer and Fowler place 4086, 4231, 
4400, 4586 and 5303 in one group; 3801, 3987 and 4567 in 
a second; and 4359 in a third. Campbell and Moore have 
three groups: 3388, 3601 and 5303; 3801, 3987 and 4567; 
and 4231 by itself in the third group. Davidson and Stratton 
grouped 3388 with 3987; 3454 with 3643; and 3601 with 
4086 and 5303. They further suggest that " 3643 should go 

1 Observatory, 54, 197, 1931. 


with 3801 and 3987," thus making one group of 3388, 3454? 
3643, 3801 and 3987. 

The 1930 slitless spectra make it very plain that coronium 
has a very uneven distribution at different position angles 
surrounding the sun. The limitations of the spectra taken 
with a slit are at once apparent. If the slit happens to fall 
on an intense part of the coronal ring, then all lines belonging 
to a group physically connected should be intensified to- 
gether; but if the slit crosses a fainter stretch of the corona, 
all lines of the group should be weakened together. Evi- 
dently the comparisons of intensities of lines at one eclipse 
with those at another (e.g., 1926 and 1929) may mean com- 
paratively little, so much depends on the position angle of 
the slit. Unfortunately, for the purpose of discussing groups, 
most of the coronal spectra up to date have been taken with 
a slit, and, moreover, it must not be overlooked that the slit- 
less spectra obtained at early eclipses have left much to be 
desired in the matter of definition. 

As already noted, the line 3601 is remarkably peculiar, 
missing in 1905 and very strong in both 1925 and 1930. In 
the last two years, the lines are very intense on both limbs 
of the sun but in a very restricted region near the sun's equa- 
tor. The line 4086 resembles 3601 very closely, and un- 
doubtedly these form a pair. With either one or the other of 
these lines the following have been grouped: 3388, 4231, 
4586 and 5303. No confirmation of this grouping is given 
by the grating spectra. 

Chief interest naturally centers in the green line at 530^. 
The 1930 spectra, showing details that in shapes resemble 
eruptive prominences, exhibit the identical details in the line 
3388 but in much weaker intensity on account of absorption 
by the speculum of the grating and the silver of the coelostat. 
The 1905 coronal spectra likewise show structural details 
similar in character in the two lines. It is interesting to find 
the two strongest lines, 3388 and 5303, unquestionably form- 
ing a pair. As already stated, 3601 and 4086 cannot be 
grouped with this pair. 

The lines 3454 and 3643 certainly do not form a pair, the 
latter being missing on Grotrian's 1929 spectra while the 


former line is strong. Mitchell's 1925 spectra likewise show 
3643 missing or doubtful. There seems little reason for 
grouping 3643 with 3801 and 3987, as suggested by Stratton 
and Davidson. If the three lines form a group, their relative 
intensities must increase and decrease together in different 
eclipses. The grating spectra and also those of Dyson and 
of Lewis do not confirm this grouping. 

All previous authorities have made a group of 3801, 3987 
and 4567. The relative intensities of the three with prismatic 
spectra have been observed as follows: (1896) 3, 5, 8; 
(1898) 3, 5, 3; (1900, 1901, 1905, Dyson) 3, 5, 6; (1908) 
3 7 4, invisible; (1926) i, 5, 3. The relative intensities vary 
so enormously from eclipse to eclipse that it does not look rea- 
sonable to combine all these or to place any combination of 
these three lines into pairs. The conclusion is confirmed by 
the slitless grating spectra. After examining all lines of the 
corona there seem to be only two pairs of lines which appear 
to show similarities in form, namely, 3388 and 5303, and 
3601 and 4087. 

All observers who have discussed slitless coronal spectra 
have come to the conclusion that the green ring at 5303 is 
weakest near the sun's axis and strongest in the sun-spot 
zones. A summary of the information 1 including the 1930 
eclipse shows that the maximum intensity is always found 
near prominences but does not necessarily coincide with the 

In comparing the images of 5303 and 6374 in the 1930 
spectra, it was surprising to find how little they resembled 
each other in their structural details of greatest strength and 
how little either resembled the high-level lines of K and Ha 
of the chromosphere. Some similarities can be detected 
between the faint streamers in the lines 5303 and 6374 and 
those in Marriott's direct photographs, but in the stronger 
details few likenesses can be found. It was further found 
that the radiation of 6374 always sticks close to the sun's 
edge and is more concentrated and more uniform than 5303. 
The details referred to above are found in the drawings re- 
produced at pages 372 and 373. From these comparisons it 

1 Astro physical Journal, 75, 21, 1932. 








U 2 1 

M <U 

/vT ~& 

u 2 




S -" 


P m 

S &C 

O H 


ffi <U 

H rC 

8 S 

^ -a- 

"t^ O 

to S 
'C P 

i 1 ."ti 



i ^ 
s g 

o ^ 

tl -^ 

o3 ^ 



Cd 4- 

o c 
w o 

43 '^ 

c: -rj 


Pu ^bO 



is evident that 5303 and 6374 cannot take their origin in the 
same atom, or at least not in the same atom in the same state 
of ionization. 

On account of the apparent simplicity of the spectrum 
of coronium, Nicholson, in 1911, began a number of investi- 
gations of great interest, the results of which have appeared 
in the Monthly Notices of the Royal Astronomical Society. 
An attempt was made by him to connect the wave-lengths 
by series relationships, and furthermore to find an explana- 
tion for these series on the hypothesis that the spectral lines 
took their origin in an atom consisting of a heavy nucleus 
surrounded by negatively charged electrons. Nicholson's 
work was the first attempt to explain spectral series by 
means of Planck's quantum theory of radiation, according 
to which interchange of energy between systems of a periodic 
kind can take place only in certain definite amounts, or 
quanta, determined by the frequencies of the systems. With 
the enormous increase in knowledge of atomic structure in 
the past twenty years, Nicholson's investigations have for 
us now merely a historical or academic interest. 

Pannekoek 1 has assumed that the coronal lines may be 
due to Ca + -f . At present we know so little of the spectral 
lines which take their origin in the singly-enhanced, doubly- 
or triply-enhanced elements, that it is quite futile to guess 
which of the many possible elements will give the particular 
series of lines seen in the visible portion of the coronal spec- 
trum. One guess is just about as good as another, and 
therefore we shall refrain from hazarding a conjecture. The 
only possible method of securing real progress is to improve 
the observational data as much as possible. 

Freeman " believes that argon is the cause of the coronal 
lines. He finds coincidences in wave-lengths for two-thirds 
of the forty lines in Campbell and Moore's list (1918). 
Many of the lines identified by him are no longer regarded 
as coronal in origin. Russell and Bowen 3 come to the con- 
clusion " that the attribution of the coronal lines to argon 
is without foundation." This is unquestionably another 

1 B. A. N., 14, 1922. :{ A^tro physical Journal, 68, 177, 1928. 

' 2 Astrophysical Journal, 69, 179, 1929. 


case, so frequent in the history of astrophysics, of the identifi- 
cation of origin from mere coincidences in wave-length. 

Hylleraas L criticizes the suggestion by Rosenthal that the 
coronal spectrum is due to helium, and finds there is no 
quantitative agreement to support this theory. 

Hopfield " finds in the laboratory spectrum of neutral oxy- 
gen a line with wave-length 6374.29 and also two lines at 
6300 and 6364 which agree in wave-length, within limits of 
error, with unidentified lines in nebulae. Several workers 
have produced the green auroral line at 5577 in the labora- 
tory. There is a remarkable agreement in wave-length be- 
tween the laboratory line and the coronal line at 6374.28. 

More recently, de Bruin 3 calculated three additional lines 
in the neutral oxygen spectrum with wave-lengths 5302.70, 
6704.07 and 6775.90. He assumes that these coincide in 
wave-length with coronal lines and hence concludes that " the 
mysterious coronium turns out to be neutral oxygen/ 7 For 
the 5303 line, the best value for the coronal wave-length 
comes from taking the mean of the 1918 and 1930 eclipse re- 
sults, which gives 5302.92; a difference too large to permit 
identifying the oxygen wave-length with the coronal. Still 
more recently, Frerichs and Dingle in Nature (129, 901, 
1932) independently discuss de Bruin's work. The former 
measures the wave-length 6374.292 for the oxygen line. 
Both of these physicists come to the conclusion that de 
Bruin's coincidences are merely accidental and that " the 
great mystery of the coronal lines remains unsolved." 

In addition to the bright line spectrum, the corona shows 
a continuous spectrum in the inner corona, 8' or 10' deep, 
with Fraunhofer absorption lines visible in the middle and 
outer corona. The early observations seemed to indicate 
that at sun-spot minimum the green coronium line was weak 
and the Fraunhofer spectrum strong, while at sun-spot maxi- 
mum the emission lines were stronger, and the dark lines 
weaker. It is only recently that observers have recognized 
that the presence of thin clouds at the time of the eclipse 
may greatly affect the visibility of the Fraunhofer lines ob- 

1 Zeit. fur Physik, 69, 361, 1931. ' 3 Nature, 129, 468, 1932. 

- Physical Review, 37, 160, 1931. 


served in the corona. At the eclipse of 1901, photographed 
by Perrine in Sumatra through thin clouds, the coronal spec- 
trum showed Fraunhofer lines in the region corresponding 
to the invisible moon! It is important, for testing various 
theories of the corona, that the distance out from the sun at 
which the absorption lines become invisible in the corona 
should be determined by adequate photographs taken under 
clear skies. 

A great increase in our knowledge of the coronal spectrum 
came as the result of the observations of the Lick Observa- 
tory L at the Australian eclipse of September 21, 1922. On 
account of the perfect weather conditions and long duration 
of totality an opportunity was afforded for carrying out an 
extensive program with exposures exceeding 5 minutes. 

The spectra taken with each of the instruments of one- 
prism dispersion exhibited the continuous spectrum and most 
of the bright lines confined to a region 4' to 6' from the sun's 
limb. The maximum was found for the green line with an 
extension of 8' from the edge of the sun. The spectrum of 
the outer corona seemed unquestionably to show the Fraun- 
hofer lines which were specially visible to the violet of Hy 
where the continuous spectrum was less intense. No trace 
of the sky spectrum was found to exist beyond the limits of 
the coronal spectrum. Since the sky was remarkably clear , 
without the slightest evidence of haze or clouds, it is manifest 
that the Fraunhofer lines of the 1922 eclipse did not take 
their origin by reflection in the earth's atmosphere. Moore's 
observations in thus proving the Fraunhofer lines of the outer 
corona are caused by the scattering of the sun's light by 
some means in the corona are of the utmost importance in 
advancing our knowledge of the perplexing solar aureole. 
Comparisons of the measures of the lines of the coronal and 
sky spectra at two points 20' east and 20' west of the sun's 
limb showed that the coronal lines on both sides of the sun 
were displaced to the red of the corresponding sky lines. The 
amount of displacement corresponds to a velocity in the line 
of sight away from the observer of 26 km per second. Since 
the light observed was reflected sunlight, the measured radial 

1 Moore, Publ. A. S. P., 35, 59, 1023. 


velocities supply evidence that the particles of the corona 
at 20' from the sun's edge are moving away from the sun 
with a speed of the order of 20 to 30 km per second. 

It is highly probable that the relative intensities of the 
emission lines of the corona vary with the sun-spot period, 
though information on this point is very meager. It is dif- 
ficult to compare spectra secured by different observers at 
different eclipses using instruments of vastly differing re- 
solving powers, especially since the coronal lines are so weak 
and are seen projected on a background of continuous spec- 
trum. Under these conditions, the intensities of the bright 
line spectrum secured with low dispersion instruments will 
obviously be less intense than for instruments of high dis- 

The spectra secured with the slit spectrographs and also by 
means of an instrument without slit seemed to prove conclu- 
sively that the coronal emission lines in 1922 were much 
fainter than those of the eclipse of 1918, also obtained by 
the Lick-Crocker expedition. The corona of 1922 was of 
the sun-spot minimum type. Hence, the suspicions of pre- 
vious observers that the bright lines of the corona are fainter 
at sun-spot minimum than they are at sun-spot maximum 
seem to be completely confirmed. 

Various observers have investigated the rotation of the 
corona by measuring from spectra the Doppler effect of mo- 
tion in the line of sight. It has been generally assumed that 
the corona rotates with the sun, which is at a rate of 2.0 km 
per second at the limb of the sun. A rotational speed of 
this size corresponds to a shift in wave-length of the coro- 
nium lines amounting to 0.035 A. And yet, the wave-lengths 
for this line in 1918 by the Lick and Mt. Wilson observers, 
each using an efficient instrument, differ by 0.2 A, or six 
times the rotational shift! It is probably no very great ex- 
aggeration to say that at the present time we know abso- 
lutely nothing regarding the rotation of the corona. It is 
not impossible that the wave-lengths of the lines 3388 and 
5303 may be determined with greatly increased accuracy, 
but these lines at best are faint, and the exposures available 
are very short. The best method of determining the rota- 


tional speed will be unquestionably by securing as accurate 
wave-lengths of these lines as possible, with the slit of the 
spectograph stretching across the sun's eqtiator. If different 
values of wave-length are obtained for east and west limbs, 
thus indicating rotation, the observer should be very careful 
that he has eliminated all possible instrumental sources. In 
place of using the solar spectrum as a comparison, it is prob- 
able that more accurate wave-lengths will be obtained by 
using an artificial source, in the manner employed with 
stellar spectra. The comparison spectrum could readily be 
superposed on the black moon. 

Wright and Curtis describe ] in detail their unsuccessful 
attempts, made at the eclipses of 1923, 1925, 1926 and 1929, 
to determine the rotation of the corona by the method ap- 
plied to the Orion nebula by Buisson, Fabry and Bourget. 

In spite of the pitifully small amount of time available 
for the investigation of the spectrum of the corona, much 
knowledge has been acquired; but, unfortunately, suspicion 
has also been cast on some of the information we thought 
was fully secured. A brief summary of what we now think 
we know may not be out of place. Three types of spectra 
must be distinguished: the continuous spectrum close to the 
edge of the sun, the emission spectrum, and the Fraunhofer 
spectrum. The coronal spectrum is now fully distinguished 
from that of the chromosphere. The lines of calcium and 
hydrogen, frequently photographed at mid-totality, do not 
belong to the corona, but are caused by the light of the 
chromosphere being diffused in the earth's atmosphere, 
usually by thin clouds. 

In order to ascertain the atomic origin of coronium, coro- 
nal spectroscopic work needs more information about (i) 
wave-lengths of greater precision, and (2 ) detection of simi- 
larities in structure. To attain the highest accuracy in wave- 
lengths a slit is desirable. The most useful form is a three- 
prism spectrograph designed to gather as much light as 
possible. To secure a wide range of wave-lengths, optical 
parts of quartz are necessary. To detect similarities in the 
structure of coronal lines slitless spectra are needed. All 

1 Sprout Observatory Publications, No. n, 1930. 


coronal investigations require large dispersion in order to 
increase the contrast between the bright-line coronal spec- 
trum and the continuous spectrum. The great difficulty is 
the pitifully small amount of light available together with 
the short exposures permitted. Unfortunately, a slit cannot 
be used with a concave grating there is not enough light. 
Without a slit there is little hope of finding detailed structure 
in the coronal lines unless the sun is in an active condition. 

Out of eight eclipses observed, the author has actually seen 
seven coronas, the 1927 phenomenon being blotted out by 
clouds. The corona that exhibited the most spectacular 
beauty and made the most lasting impression was not the 
first eclipse witnessed, that of 1900, but the eclipse of 1918. 
This was the eclipse of pronounced color, the prominences 
were large and brilliant. The contrast between the warm, 
rosy color of the prominences and the pearly-white filmy 
structure of the corona left a never-to-be-forgotten impres- 
sion on the mind, that time can never efface. On account of 
the greater number of prominences visible at sun-spot maxi- 
mum, the eclipse near maximum provides a more wonderful 
picture to the unaided eye than does the minimum type of 
corona. The eclipse of 1927 therefore presented to those 
fortunate enough to witness it, as gorgeous a spectacle as 
that of the eclipse of 1842 (see p. 132 ) which aroused both 
the populace and the astronomers to such a high pitch of 
enthusiasm that every eclipse from that day to this has been 
assiduously observed. 

To observe and enjoy the beauty of the corona one needs 
little but his own good eyes. A good pair of field glasses or 
a small telescope will permit the study of the prominences or 
of the details of the inner corona. Baily's beads, at the be- 
ginning and at the end of totality can be more fully enjoyed 
by the use of telescopic power. It should hardly be neces- 
sary to add that one should protect the eyes from the glare 
of the sun while the crescent sun is diminishing before the 
advent of totality. 



AS THE Fraunhofer lines undoubtedly exist in the 
spectrum of the outer corona, they must be caused by 
the reflection of chromospheric light by matter exist- 
ing in the corona in a finely-divided state. Fortunately, there 
are methods available for testing the question of scattering 
in the corona, namely, by observations for the determination 
of the polarization of light. Since the eclipse of i860, 1 when 
Secchi and Prazmowski first took up the subject, polariza- 
tion observations have found their place at almost every 
total eclipse. At the beginning of the investigations, they 
were carried out entirely by visual methods; but in a manner 
similar to what has happened in other branches of astro- 
nomical work, photographic observations have gradually 
displaced visual ones, and as a consequence greater and 
greater precision has been attained. The accurate determi- 
nation of the percentage of polarized light has an important 
bearing on the study of the distribution of matter in the 
corona. To be of the greatest value, the polarization should 
be known for different distances from the sun's limb. 

In general, there are two different methods of analyzing 
polarization; one is by the use of double-image prisms, the 
other by means of plane mirrors. There are many varia- 
tions of these methods possible. In 1900, Wood employed 
a direct vision prism before the object-glass of a telescope, 
the eye-piece containing a Savart plate and a Nicol prism. 
This combination gave a continuous spectrum crossed by 
very distinct diagonal interference bands, manifesting fairly 
strong polarization estimated to equal between 10 and 15 
percent. The character of the interference bands indicated 
that the bright-line spectrum was not polarized, or in other 

1 Comptes Rendus, 51, 195, 1860. 



words, that the light causing these lines was not reflected 
sunlight. The appearance presented in the telescope how- 
ever differed so materially l from what had been expected 
that it took many of the precious seconds of totality for the 
observer to readjust his ideas; and all the while he could not 
help but feel that something radically wrong must have 
happened to the apparatus. At the same eclipse, Dorsey 2 
photographed the corona through a double-image prism in 
the manner utilized by A. W. Wright in 1878. The method 
gives two photographs of the corona on each plate; one 
having cut out of it all the light polarized along the line 
joining the two images, and the other all that polarized at 
right angles to this direction. Hence, if the corona is 
polarized radially or tangentially, one image will be deficient 
in light along the diameter perpendicular to this direction. 
Which image is deficient along the line joining the centers 
of the two photographic images depends on the kind of 
double-image prism used and whether the polarization is 
radial or tangential. Dorsey also examined the corona 
visually by a polarimeter consisting of a telescope in the 
focal plane of which was placed a biquartz half an inch 
square. The eye-piece contained a Nicol prism, and be- 
tween the objective and the biquartz was a double pile of 
plates. The conclusions were that the corona is polarized 
radially, the visual observations giving the amount of eleven 
per cent at a distance of 8' from the moon's limb, a value 
agreeing well with Wright's 11.2 per cent found at f from 
the moon's limb at the eclipse of 1878. 

No attempt can be made to give here a complete account 
of the numerous observations made at various eclipses to 
determine the amount of polarization. Special mention, 
however, should be made of the excellent work by Newall 
and by Turner, each of whom has observed polarization 
effects at several eclipses. In Lick Observatory Bulletin, 6, 
1 66, 1911, R. K. Young discusses the measures of photo- 
graphs secured by Perrine of the Lick Observatory at 
eclipses of 1901, 1905 and 1908. At the Sumatra eclipse, 

1 Publications of the U. S. Naval Observatory, 4, D 116, IQO$. 

2 Publications of the U. S. Naval Observatory, 4, D 117, 1905. 


Note the difference in the distribution of coronal light in the two photographs. 


Measures by Bcrgstrand in units of stellar magnitude. 


the plates were secured by a double-image camera with the 
prism set in succession at five different positions separated 
by angles of a quarter of a right angle. At the two succeed- 
ing eclipses, photographs were secured by the same double- 
image camera and also by a reflecting polarigraph. This 
latter consisted of three cameras with lenses of three inches 
aperture and fifty inches focal length. In front of each of 
two of the lenses was placed a glass reflector, so inclined 
that the light from the corona was incident at the polarizing 
angle, the planes of polarization of the two reflectors being 
perpendicular to each other. The third camera was used 
merely as a check. The measures showed that the polariza- 
tion was radial and that the percentage of observed light 
increased rapidly from the limb, reaching a maximum of 
thirty-seven percent at 5' distance, and then diminished 
slowly, being thirty-five percent at 9' from the limb. As- 
suming the well-known law of the reflection and scattering 
of light that it varies inversely as the fourth power of the 
wave-length, the value of u per cent in the visual re- 
gion 5600 A would correspond to 33 per cent in the photo- 
graphic region 4270 A. A close accord is thus seen to exist 
between the visual values obtained from the eclipses be- 
fore 1901, and the photographic results from the three 
eclipses of 1901, 1905 and 1908. In view of the very 
great difference between the amounts of polarization in the 
visual and photographic regions, it is highly desirable that 
values in the visual region be obtained by photographic 
methods by the use of a color filter and isochromatic plates. 
In view of the experience of Wood at the eclipse of 1900, 
it is urged that all observations in the future for polar- 
ization effects be made photographically. Observations 
might possibly still be made visually by experienced 
observers, but such values should be looked upon merely 
as checks on the more accurate photographic results. Ob- 
servers should be most careful to know the amount of polar- 
ization caused by the apparatus itself so as to eliminate 
these effects from the total polarization observed during 
the progress of the total eclipse. As is well known/ every 

1 Wood, Astrophyskal Journal, 12, 283, 1900. 


form of apparatus that disperses light, at the same time po- 
larizes it. A Rowland grating gives strongly polarized spec- 
tra; and with prismatic spectra, as the dispersion is in- 
creased by additional prisms, the polarization is likewise 
increased by the new surfaces added. 

At the eclipse of 1918, Lewis of the Lick Observatory 
party, by means of two separate double-image cameras, se- 
cured successful photographs in two different regions of the 
spectrum by using blue and green color filters. The effect 
for the blue was found to be greater than for the green. 
Quantitative values for the amount of polarization could not 
be furnished however for two reasons: first, the law of 
diminution of the intensity of coronal radiation at different 
distances out from the moon's limb is unknown; and second, 
the effect on the corona, of polarization of the light of the 
sky surrounding the corona, has not been fully investigated. 
At the eclipse of 1905, Newall ' found that at a distance of 
three-quarters of a degree from the center of the corona, 
the strength of the Savart bands from the sky neutralized 
those from the corona. This signifies that from the veil of 
the illuminated sky between the observer and the corona 
there came as much polarized light as from the corona three- 
quarters of a degree from the center. Unquestionably, the 
character and intensity of the atmospheric polarization vary 
considerably at different eclipses, which of necessity are 
observed under different conditions of clouds and moisture 
in the terrestrial atmosphere. On account of the great in- 
tensity of the corona close to the sun, for instance, at i' 
from the limb, it is difficult to measure the intensity of the 
darkening of the photographs and hence to evaluate the 
amount of polarization so close to the edge of the sun. 
Newall in 1901 obtained " quite marked polarization " at 
i' from the limb. The Savart photographs for testing po- 
larization seem to possess some advantages over the double- 
image or reflection methods (see Newall, loc. cit.). 

The only means of unravelling some of these puzzles 
seems to lie in determining the law of change in the intensity 
of the corona at different distances out from the limb of 

i Monthly Notices, R. A. S., 66, 475, 1906. 


the sun. The law best known is that of Turner, 1 as the re- 
sult of photographs obtained in 1898, that the intensity of 
the corona varies from the edge of the sun outwards in- 
versely as the sixth power of the distance measured from the 
center of the sun. At the eclipse of 1905, Schwarzschild L> 
confirmed Turner's law, and this same eclipse, Graff d as- 
sumed the correctness of this law to determine the law of 
blackening of his photographic plates. But at this same 
eclipse of 1905, Becker 4 found the intensity of the corona 
subject to a different law, that it varied inversely as the 
fourth power of the distance counted from a point one- 
seventh of a solar radius inside the edge of the sun. At the 
eclipse of 1908 by means of measures carried out by the 
bolometer, Abbot r> confirmed Becker's law rather than that 
of Turner, while R. K. Young (loc. cit.) found an intensity 
depending on the inverse sixth and eighth powers of the 
distance measured from the center of the sun. 

A very different law was found by Bergstrand in a very 
important publication entitled " Etudes sur la distribution 
de la lumiere dans la couronne solaire," Upsala, 1919. From 
photographs secured at the eclipse of August 21, 1914, an 
attempt was made to determine the relative intensity of 
light distributed within the corona. The measurement of 
the absolute intensity and the estimation of the total light 
of the corona, compared for instance with that of the full 
moon, did not form part of the program. The problem is 
one of photometry, and for its solution can be brought the 
vast experience gained by many years of investigation in 
determining the magnitudes of the stars. Of the several 
methods available, Bergstrand adopted the plan of employ- 
ing twin photographic objectives, mounted equatorially in 
such a manner that the two solar images could be impressed 
upon one and the same photographic plate. On the day of 
the eclipse the times of exposure of the two objectives were 

1 Popular Astronomy, 14, 548, 1906. 

2 Astron. Mitteil. zu Gottingen, 13, iqo6. 

3 Astron. Abhandl. der Hamburger Sternw. in Bergedorf, j, i, 1913. 

4 Memoirs, R. A. S., 57 and Phil. Trans. Roy. Soc. 207 A, 1908. 
6 The Sun, 133, 1911. 


made identical, but the aperture of one of the objectives 
was reduced by means of a suitable diaphragm to one-third 
that of the other. 

The intensity of the silver deposit measured on the plates 
is the summation of two separate effects, one of which is 
due to the corona itself while the other comes from the dif- 
fuse light of the sky. Added to these two, there is in reality 
a third effect found close to the moon's limb, that of a halo 
caused by reflection from the glass-side of the plate of the 
strong illumination of the inner corona. Fortunately, the 
intensity of the corona could be separated from the two 
other effects. The values thus secured do not in any man- 
ner confirm Turner's law of the inverse sixth power nor yet 
the law of Becker according to which the intensity varies in- 
versely proportional to the fourth power. In fact, Bergstrand 
finds that the intensities near the solar equator differ greatly 
from those near the poles, the equatorial rays having an in- 
tensity three times as great as the polar rays. The equa- 
torial and polar intensities, however, can be brought into 
relationship with each other in a very simple manner by 
supposing that the corona is composed of two phenomena. 
One of them, the " interior corona " exists exclusively in the 
equatorial zone. In both of these phenomena, the intensity 
of the light is inversely proportional to the square of the 
distance measured from the edge of the sun, the intensity 
of the " equatorial corona " being, however, about double 
that of the " interior corona." On page 389 is given a curve 
representing Bergstrand J s values. Measures were carried 
out on solar radii separated from each other by 15. Posi- 
tion angles are designated from the north towards the east. 
The intensity at a distance of one radius from the edge of 
the sun is taken as unity, and values are represented in terms 
of stellar magnitudes, where a difference of five magnitudes 
represents a change of a hundred-fold intensity. The 
strongest coronal rays at times depart sensibly from the 
direction of the solar radius. This is shown by a jet which 
leaves the sun at position angle 25 but which does not go 
out radially and is found, in the external curves between 
30 and 45. Some of the most intense rays apparently 


do not take their origin from the edge of the sun but rather 
from the front or back side of the solar disk. Moreover, 
the structure of the corona is highly complicated, since the 
distribution of jets is not uniformly distributed in all longi- 
tudes and since they frequently depart sensibly from the 
radial direction. On account of the greater strength of the 
equatorial rays, it was possible for Bergstrand to observe 
these rays on the photographic plates to a distance of ten 
radii, or five solar diameters from the edge of the sun, before 
they diminish in intensity to that of the diffuse sky light. In 
the polar direction, equality was attained at a distance of 
three and a half solar diameters. 

Valuable observations were secured by photography at the 
1925 eclipse 1 by Pettit and Nicholson. They used a Ross 
6-inch doublet of 1 5-foot focus fed by a coelostat. Exposures 
were made in photographic light and also at wave-lengths 
greater than 6100 A. Standard photometric squares were 
impressed on each plate. The plates were mounted on a 
turntable and measures were made by the Koch micropho- 
tometer. Two corrections were applied to the measures, one 
on account of general scattered sky light, when it was as- 
sumed that the mean density of the plate at the four cor- 
ners gave a measure of the sky illumination. The other cor- 
rection was for halation and scattered coronal light. No 
correction was applied for the polarization at the coelostat 
mirror. Pettit and Nicholson find that their measures in the 
photographic region confirm the inverse sixth power from the 
center of the sun, while those in the visual region obey the in- 
verse seventh power. However, the measured values both 
in the photographic and visual regions more closely con- 
form "' to the inverse fourth power measured from a point not 
coinciding with the center of the sun. The 1925 photographs 
were measured to a distance of three radii from the sun's 
center while Bergstrand 's 1914 photographs were measured 
to ten radii. Also at the 1925 eclipse, King and Miss Har- 
wood 8 found that the intensity of coronal radiation at dis- 

1 Astrophysical Journal, 62, 202, 1925. 

2 Handbuch der Astro physik, 4, 333, 1929. 

3 Harvard Circulars, 312, 1927. 


tances from the sun's limb greater than io'.$ follows the law 
of inverse squares. 

Following the eclipse of 1926, Stetson and Andrews ! dis- 
cuss the measures made by different observers (including 
themselves) and come to the conclusion that no single law 
can express the brightness of the solar corona as a simple 
function of the distances either from the limb or from the 
sun's center. Within 2% radii of the center, the intensity 
appears to decrease as the seventh power, from 2-\ to 3 radii 
the inverse fourth power fits better, while further out the law 
is the inverse square. 

At the 1927 eclipse, Belanovsky and Perepelkin J secured 
photographs through thin clouds which were measured by 
the Hartmann microphotometer. From the measures, lines 
of equal intensities, or isophotes, were drawn which showed 
that the corona decreased in intensity according to power 2.7 
with respect to distances taken near the edge of the sun. 

At the 1929 eclipse, the German expedition secured ' ex- 
cellent photographs with a horizontal camera of 28-foot focus 
and with an astrographic telescope of i i-foot focus. A thor- 
ough discussion has been made by von Kliiber. The photo- 
graphs of both cameras were measured by the Hartmann 
microphotometer and isophotes were drawn giving lines of 
equal intensity in the corona. The astrographic plates show 
that the corona decreases in intensity according to power 2.5 
in units of distances from the edge of the sun. With the hori- 
zontal camera, at the sun's poles the plates show that the 
decrease in coronal intensity is according to the power 2.0, 
while at the equator the exponent 2 .3 more correctly repre- 
sents the observations. All of the 1929 measures together 
are best represented by the inverse power 2 .4 times the dis- 
tance from the sun's edge. 

It is evident that our knowledge regarding the distribu- 
tion of light within the corona is in a very unsatisfactory 
state, since the law of the intensity has been found to be 
inversely as the second, fourth, sixth or even eighth power 

1 Astro physical Journal, 6g, 227, 1925 

- Monthly Notices, R. A. S., 88, 740, 1928. 

>A Zeit. fur Astrophysik, 2, 289; 3, 142, 1931. 


of the distance from the sun. It is consequently of great 
importance that a well-devised form of apparatus be con- 
structed for use at eclipses, and that photographs be secured 
both in the violet and visual regions on a carefully prepared 
plan at several future eclipses. The apparatus used by 
Bergstrand seems to leave little room for improvement. 
If the photographic plates secured at the eclipse could be 
impressed by light from a standard source, and if in addition 
photographs of the full moon were obtained, we should then 
be in a position of having information additional to that 
acquired during the progress of an eclipse. We need to 
know whether the intensity of the distribution of light 
within the corona follows the same law at every eclipse, or 
whether this law varies according to the sun-spot period, 
and we need to know the law both in the blue and yellow 
regions. Many of the coronas have been observed through 
clouds or haze, or with varying conditions of transparency. 
Unfortunately it has not been possible to make proper al- 
lowances for these varying factors with the result that the 
observations are undoubtedly affected by systematic errors. 
Not until we secure more and better standardized observa- 
tions can we expect to advance much in the solution of coro- 
nal problems. 

In considering the measurements made for determining 
the total light of the corona, we see similar evidences of large 
systematic errors depending on transparency conditions at 
the time of the eclipse. The following table is taken from 
Handbuch der Astrophysik, 4, 336, 1929, and brought up to 
date. The total light of the corona is expressed in terms of 
the total light from the full moon. 

The photographic determinations at the eclipse of 1886 
and the two eclipses in 1889 give values that are apparently 
too small, while one value in 1898 is too large. Giving half 
weight to each of these four determinations, the weighted 
mean shows the total light of the corona to possess 55 per 
cent of the light of the full moon. Results from one ex- 
pedition may be given here. Kunz and Stebbins observed 
the 1918 eclipse with a potassium photo-electric cell. They 
compared the light of the corona with a standard candle, 







In terms of 
full moon 



W, H. Pickering 


1889, January 


o 04 

1889, December 





o b 



r i 


Bacon and Gare 

2 7 



O 20 



o 17 






Abney and Thorpe 

o 8 

1889, January 


o 4 


Abney and Thorpe 

I 1 



o 75 



o 85 




o 20 



Kunz and Stcbbins 

o 50 




o 41 




o 27 



Petti t and Nicholson 

o 52 



Stetson and Coblentz 

o 52 

with two electric lamps, with the full moon, and with an area 
of the sky during totality and during full sunshine. Their 
numerical results are as follows: 

Observed total light of the corona 0.60 candle-meters 
Same, corrected to outride of atmosphere i 07 candle-meters 
Observed ratio of corona to full moon o 6 
Same, corrected to outside of atmosphere 0.50 
Observed ratio of corona to sky circle of diam- 
eter one-half degree and 8 from uneclipsed sun 0.105 
Same during totality 640. 

The photo-electric cell with which these measures were 
secured had a maximum sensibility at wave-length 4500 A 
in the blue. 

Further details about the earlier eclipses may be obtained 
from Handbuch der Astrophysik. In order to illustrate the 
inconsistencies in the best work along these lines, a few 
words may be said about the more recent eclipses, begin- 
ning with 1925. At this eclipse, Harvard Observatory occu- 
pied four stations in order to measure the total light of the 
corona by a photometer of " pinhole " type designed by 


King. It was found that the integrated brightness within 
a circle 3 in diameter, given in stellar magnitudes, is: 
photographic -10.96, photovisual 11.61; within a cir- 
cle 6 in diameter the values are -11.40 and -11.71, re- 
spectively. After eliminating the effect of the illuminated 
sky, the stellar magnitude of the corona is: photographic 
- 10.76, photovisual - 11.57. With one of these photome- 
ters at the 1926 eclipse, Stetson and his co-workers ob- 
tained conflicting results in the circles 3 and 6 in diameter; 
but the measures seemed to show that the corona was 40 
per cent brighter in 1926 than in 1925. 

It is interesting also to compare the measures of the total 
illumination of corona plus sky. In 1925 the horizontal illu- 
mination was 0.24 foot-candles, or equal to the total in- 
tensity 30 minutes after sunset; in 1926, measures by the 
Macbeth illuminometer gave 0.14 foot-candles; in 1929, a 
value 0.15 and 1930 (Niuafoou Island) 0.38 foot-candles. 
The measures on the three last eclipses were carried out by 
Stetson's instruments. The above measures seem to show 
that although the 1926 corona was 40 per cent brighter than 
that of 1925, the total illumination of corona and sky to- 
gether was 40 per cent fainter. 

Another instance of the uncertainties underlying meas- 
ures of coronal radiation was shown at the 1926 eclipse. 
With the photographic photometer it was found that the 
total brightness of the corona was but one-tenth that of 
corona plus sky as measured with the illuminometer. 

At the eclipse of 1918, Aldrich 1 found that the total 
brightness of the sky during totality was less than that of 
twilight one hour after sunset of the same day. At the 
Australian eclipse of 1922, Ross directed a camera, from 
which the lenses had been removed, toward the south 
celestial pole and exposed a photographic plate during to- 
tality. Other plates from the same box were exposed on 
the evening of the same day for equal intervals of time at 
6:14, 6:17, 6:20, 6:23 and 6:26 by the clock. The central 
portions of the plate were cut out and all six plates were 
developed together. The plates showed a regular gradation, 

1 Smithsonian Miscellaneous Collections, 60, No. 9, 1919. 


It was found that the illumination at the south celestial pole 
corresponded with that when the sun's center was 97 29' 
from the zenith. 

It is evident from the above summaries of measures that 
our present knowledge regarding the total amount of light 
received from the corona is in a very unsatisfactory state. 
Photometric methods present many complications. When 
results at one eclipse are compared with those of another, 
the instruments must be thoroughly calibrated and methods 
properly standardized before we can have any confidence in 
the conclusions. To obtain the total light of the corona, it 
is necessary to extrapolate to the edge of the moon and even 
to the edge of the sun. Unfortunately we seem to have little 
sound information about the law of intensity in the corona 
close to the sun's edge. The greatest difficulty of all is that 
the corona perforce must be observed through clouds and 
haze at half the eclipses where the corona is at all visible 
and through varying transparencies at the other half. Up 
to the present we know of no adequate methods of making 
proper allowance for varying atmospheric conditions. Tak- 
ing everything into consideration, it seems highly probable 
that the accidental and systematic errors existing in meas- 
ures of coronal radiation certainly amount to 20 per cent, 
and possibly 50 per cent, of the final values. Roughly 
speaking, the total light of the corona is one-half that of the 
full moon and one-millionth that of the noon-day sun. 

Hence we must not take too seriously the attempts to cor- 
relate intensity of coronal radiation with sun-spot activity. 
It will be necessary to await more accurate observations of 
the future. It seems entirely probable that the inner corona 
at sun-spot maximum must be brighter than at sun-spot 
minimum. Moreover, as the inner corona contributes the 
most energy to the total coronal radiation we would there- 
fore logically expect that the total energy at maximum of 
spots is greater than at minimum. But to confirm this from 
observations already secured is another story. When we 
consider the total amount of time available for coronal in- 
vestigations, we must not be too discouraged with the re- 
sults obtained to date. 


In former chapters it has been shown that for half a 
century it has been recognized that there is a close connec- 
tion between the shape of the corona and the sun-spot cycle. 
At. minimum of spots the corona shows the long equatorial 
extensions and the strong polar brushes, while at spot maxi- 
mum the corona is more nearly circular in outline. The first 
to call attention to this was Ranyard in 1879 in the " Eclipse 
Volume " of the Royal Astronomical Society. In 1897, Han- 
sky made the connection more certain by publishing a series 
of reproductions of the corona arranged according to the 
sun-spot curve. 

In Handbuch der Astrophysik, 4, 317, 1929, there is stated 
the general problem. For the past forty years, since accu- 
rate photographs became available, the dates of maxima and 
minima of spots have been as follows: 

Maximum Minimum 





The eclipses of the past decade have been beautifully 
situated with respect to the spot cycle. The eclipse of 1922 
took place shortly before minimum of spots, that of 1923 
almost exactly at minimum. A year and a half after mini- 
mum came the 1925 eclipse, while those of 1926 and 1927 
were just before and 1929 and 1930 just after maximum. 

The information on the activity of the sun, as recorded 
in the sun-spot cycle, comes, as the name signifies, from 
observations on spots, their Zurich relative numbers, the 
Greenwich means areas, the latitude of spots, etc. Similarly 
observations made on prominences, give also the number, 
mean areas, mean latitude, etc. The sun-spot and prominence 
curves closely parallel each other. In a sense, the informa- 
tion from prominences supplements that from spots in that 
the spots are phenomena observed on the face of the sun while 
the prominences are photographed only at the sun's limb. 


To observe spots a moderate equipment, merely a tele- 
scope, is all that is necessary. The activity of the sun 
manifesting itself in spots can be observed as the spot moves 
across the face of the sun; measurements of polarity, etc., 
continued from day to day, give a fairly faithful indication of 
the relative activity of the sun. In comparison, the promi- 
nences are transitory phenomena. Even if individual promi- 
nences were more permanent than they are, the rotation of 
the sun would carry them quickly out of sight. Occasionally 
when a total eclipse comes, the direct photographs of promi- 
nences and the eclipse spectra can be compared with spec- 
troheliograph photographs. Then we see the limitations of 
the latter method. The eclipse photographs on October 21, 
1930, showed the sun with a very stormy region in the south- 
east quadrant, while the Mt. Wilson spectroheliogram taken 
only a few hours earlier showed nothing particularly re- 
markable about the activity in this quadrant. Unless the 
prominence exhibits both height and contrast, through masses 
of gases in eruption, it is usually not a conspicuous object 
on the spectroheliograph plates. Hence, all things consid- 
ered, it has been generally felt by the average astrono- 
mer that the information from spots gives a more reliable 
indication of the activity of the sun than is obtainable 
from observations of prominences. As already stated, 
however, the two types of observation supplement each 

When attempts are made to find correlations between 
coronal disturbances and either spots or prominences, it is 
evident that many more connections must always be found 
between prominences and coronal structure than between 
spots and corona. The spots are seen on the face of the sun 
while both the corona and prominences stretch out from the 
edge of the sun. It must not be thought that there must be 
a more intimate connection between prominences and coronal 
activity than there is between sun-spots and coronal disturb- 
ances. The eclipse photographs show prominences and dis- 
turbed regions in the corona but from the nature of things 
cannot show spots. Unless the eclipse astronomer observes 
the spots on the final days before the eclipse and he is 

Photographed with the 63 -foot tower telescope by the Swarthmore College Expedition. 


usually too busy with a thousand and one preparations that 
must be made or unless he looks up the literature after- 
wards (and this is rarely done), the possible connections 
between individual spots and coronal disturbances pass un- 

Ludendorff has done a valuable piece of work in providing 
a simple method of measuring the shape of the corona. He 
had the happy inspiration of utilizing the published photo- 
graphs and half-tone reproductions to draw roughly by eye 
lines of equal intensities (isophotes) in the corona. In some 
instances, isophotes were already available in the original 
sources, especially in the corona of 1905 from the investi- 
gations of Graff and in that of 1914 from the hands of 
Bergstrand. Comparisons of the results from Ludendorff 's 
tracings with the more accurate methods of Graff and Berg- 
strand showed no systematic differences. From all of the 
tracings of each eclipse, two quantities a and b, were deter- 
mined by the method of least squares. The value a is the 
ellipticity of the corona at the edge of the sun, while a + b 
is the ellipticity at a distance of one radius out from the 
sun's edge. From the examination of thirteen eclipses be- 
tween 1893 an d 1927, Ludendorff found that the quantity a 
was nearly constant for all eclipses. On the contrary, the 
value b varied. For the coronas near sun-spot maximum, 
the value of b is very nearly zero, the corona being approxi- 
mately circular in outline. Near minimum of spots, how- 
ever, b has increased in value, the corona being more 

Mitchell has found a closer connection between the coro- 
nal shape and sun-spot numbers than there is with the 
phase in the spot cycle measured from the time elapsed 
from minimum or maximum of spots. The table in Hand- 
buck der Astrophysik, 4, 338, 1929, is here extended to in- 
clude the eclipses of 1922, 1929, and 1930. The values from 
the 1929 eclipse are from isophotes discussed by von Kliiber 
(loc. cit.}. Photographs of the 1922 eclipse were kindly sent 
me by Moore, and of the October, 1930 eclipse by Marriott. 
The measures of both eclipses (following Ludendorff 's 
methods) were made by Miss Williams. 












i $93 3 


o 03 

1926 o 


o 06 


igi8 4 


o 13 


1927 5 


o 04 


1929 3 




1905 7 




1908 o 


o 06 

O O2 

1030 8 


o 04 

o 23 

i8g8 i 


o 06 

O 12 

1896 6 


o 03 

o 23 

1901 4 

1 1 

o 04 

o 26 

1923 7 


o 06 



o 03 

o 29 

1914 6 


o 05 

1925 i 


o 05 

1922 7 


o 04 

o 24 

o 03* 


* All values of a and 6 are positive except b of the 1893 eclipse 

On the left side of the table are the eclipses between 
maximum and minimum of spots, or on the descending 
branch of the sun-spot curve, while on the right hand are 
the eclipses after minimum with spots increasing in numbers. 
One should read down on the left and then up on the right 
side with spots increasing. Instead of arranging the ma- 
terial according to the phase in the spot cycle, as was done 
by Ludendorff, it has been arranged according to the mean 
of the spot numbers in a synodic month of the sun's rota- 
tion, i.e., the mean of 13 days before and after the eclipse, 
also the day of the eclipse itself. It might be added that 
von Kliiber, from isophotes drawn from microphotometer 
measures of the 1926 eclipse, obtained + 0.05 and +0.02 
for a and b respectively, in excellent agreement with Luden- 
dorff 's rougher measures. 

Bergstrand * also takes Ludendorff 's figures and applies a 
correction to the values of the isophotes in polar regions to 
allow for the effect of equatorial and mid-latitude streamers 
being superposed over polar rifts. He finds a correlation 
with sun-spot numbers but a closer connection with promi- 
nences in high latitude zones. 

Lockyer z makes a very valuable compilation by collecting 

1 Arkiv. Mat. Astr. octi Physik. A, 22, No. i, 1930. 
- Monthly Notices, R. A. S. 91, 797, 1931. 


into one diagram, beginning with the year 1860, information 
regarding sun-spots, prominences and coronal shapes. 

On account of the connection between coronal and other 
solar disturbances, already alluded to, and on account of the 
fact that prominences are much more widely distributed in 
heliographic latitudes than are spots, it is not surprising that 
Lockyer finds a close connection between the shape of the 
corona and prominences, the maximum type of corona oc- 
curring when prominences are near the sun's poles. 

The quantities a and b in the table are exceedingly inter- 
esting. The mean of all 16 values of a is 0.05. The only 
eclipse with outstanding values, that of 1918, is explained in 
Handbuch der Astrophysik. 

The three eclipses 1900, 1901 and 1922 took place on the 
average 0.9 years before the time of minimum of spots. 
Each of these eclipses showed the typical " minimum " co- 
rona; the mean value of b is 0.26. On the other hand, the 
three eclipses of 1914, 1923 and 1925 took place the same 
interval (0.9 years) after spot minimum, and yet the coronas 
had lost their minimum characteristics, b having the mean 
value 0.15. The three eclipses 1896, 1898 and 1930 took 
place about four years before minimum of spots. The mean 
value of b amounting to 0.19 is greater in size than that at 
a time less than one year after minimum. Hence, it is evi- 
dent that two years before spots and prominences are at a 
minimum, the corona takes on the " minimum " shape. 

With the recent spot maximum occurring at 1928.5, the 
eclipses of 1926, 1927 and 1929 were all " maximum-type " 
or circular coronas. The 1926 eclipse took place 2^ years 
before spot maximum. It would seem also that the " maxi- 
mum " type of corona takes place two years before maximum 
of spots. The greatest surprise in the whole of the tabular 
values is for the eclipse of October, 1930. Taking place a 
year and a half after the 1929 eclipse and only a little more 
than two years after maximum of spots and probably four 
years before the next minimum, the value of b was approxi- 
mately equal to that of a minimum-type of eclipse. 

Bernheimer l agrees with the writer that the minimum 

1 Meddel. Ltinds Astron. Obs. f Ser. I, No. 126, 1931. 


type of corona takes place before minimum of spots. He 
also finds that the coronal shapes are more closely connected 
with sun-spot numbers than with the phase in the spot cycle. 
After observations of sun-spots during three hundred years 
we are forced to the conclusion that the n-year curve must 
be regarded as quite erratic. The solar activity from num- 
bers and areas of spots, or from frequency and areas of 
prominences, gives us curves which are similar but are not 
identical. Other closely allied curves, such as terrestrial mag- 
netic phenomena, at times depart from the general run of 
the spot and prominence curves. In spite of numberless 
investigations, it is still unsafe (see p. 120) to predict, even 
a few years in advance, what the sun-spot cycle is going to 
do. Recognizing these limitations, or in other words, realiz- 
ing that the sun has sporadic and unexpected bursts of ac- 
tivity, it is well not to make too great claims for the de- 
pendence of anything on the sun-spot curve. 

The corona takes its shape primarily from the lengths and 
position angles of the longest streamers. The eclipse of 
1930 showed that the longest streamer was connected with 
solar activity that had persisted for several days. Hence it 
appears certain that the corona can have no constant shape 
and that no doubt it varies from day to day depending on 
whether the active solar areas are near the sun's limb when 
only they could give an effect projected on the sky back- 
ground. The corona we happen to see or to photograph 
during the few fleeting seconds of totality is a temporary 
phenomenon which will subject itself to exact analysis with 
great reluctance. 

In spite of all that is said above, the author is going to be 
rash enough to predict that the 1932 corona, two years be- 
fore the expected minimum of spots, will show the minimum 
type of corona with long equatorial streamers and strong 
polar brushes. 

When we attempt to solve the enigma of the corona, we 
are face to face with one of the most difficult problems in 
the whole realm of astronomy. 

First of all, the corona is not an atmosphere of the sun 
consisting of atoms and molecules attracted by gravity. But 


gravity on the sun is 2 7 times more powerful than the value 
on the earth, while the corona has been observed to the 
enormous distance of ten million miles from the surface of 
the sun. Thus it is easy to see that if the corona were 
truly atmospheric in its nature, the resulting pressure would 
be colossal. The chromosphere is indeed the solar atmos- 
phere, not one of oxygen and nitrogen as on the earth, but 
of the gases found to exist in the flash spectrum, each heated 
to high temperatures due to their proximity to the sun, and 
yet the pressure at the base of the chromosphere is much 
less than one-thousandth of the earth's pressure at sea- 

In the forceful words of Simon Newcomb 1 we may re- 
mind the reader that " the great comet of 1843 passed with- 
in three or four minutes of the surface of the sun, and there- 
fore directly through the midst of the corona. At the time 
of nearest approach its velocity was 350 miles per second, 
and it went with nearly this velocity through at least 300,000 
miles of corona, coming out without having suffered any 
visible damage or retardation. To form an idea of what 
would have become of it had it encountered the rarest con- 
ceivable atmosphere, we have only to reflect that shooting 
stars are instantly and completely vaporized by the heat 
caused by their encounter with our atmosphere at heights 
of from 50 to 100 miles; that is, at a height where the at- 
mosphere entirely ceases to reflect the light of the sun. The 
velocity of shooting stars is from 20 to 40 miles per second. 
Remembering, now, that resistance and heat increase at 
least as the square of the velocity, what would be the fate 
of a body, or a collection of bodies like a comet, passing 
through several hundred thousand miles of the rarest at- 
mosphere at a rate of over 300 miles a second? And how 
rare must such an atmosphere be, when the comet passes 
not only without destruction, but without losing any sensible 
velocity? Certainly so rare as to be entirely invisible, and 
incapable of producing any physical effect.'' Other comets, 
the great comet of 1882 for instance, have almost grazed 
the sun's surface. 

1 Popular Astronomy, 265, sixth edition, 1887. 


Any adequate theory of the corona must be capable of 
explaining 1 the following facts: 

1. The total brightness of the corona is very small, being 
about one-half that of full moon and one-millionth that of 
the sun. Half of the total light comes from a zone extending 
only 3' from the limb of the sun. 

2. Its spectrum shows the bright emission lines of co- 
ronium extending to a maximum distance of 8', and also a 
continuous spectrum in the inner corona and Fraunhofer 
lines in the middle and outer corona. 

3. The emission lines of the spectrum are fainter at mini- 
mum of spots than at maximum. 

4. The emission spectrum due to " coronium " contrib- 
utes only a small fraction of the total energy of coronal 
light. The predominant part lies in the continuous spec- 
trum. The distribution of energy in this continuous spec- 
trum differs little from that of the sun. 

5. Polarization is a maximum at a distance of about 5' 
from the sun's limb, and it diminishes more rapidly towards 
the sun than away from it. 

6. Matter of any kind so close to the sun must be very 
hot and must reflect and scatter the solar rays. 

7. According to the observations of Abbot, the coronal 
materials are deficient in heat rays. 

8. If the sizes of the coronal particles change in diameter 
at different distances from the sun's limb, a corresponding 
change in color of the corona would result. No change in 
color is noticeable. 

9. The internal motions in the corona are very small. 

10. The sun exhibits a magnetic field. 

11. According to Bergstrand, the intensity of the corona 
varies inversely as the square of the distances from the sun's 
surface. Other authorities derive the inverse fourth, sixth, 
seventh or eighth powers. 

12. It is necessary to explain the changing form of the 
corona with variation in the sun-spot period. 

In the bright inner corona is found the emission spectrum 

1 See also Abbot, Smithsonian Misc. Collections, 52, 31, 1908, and Lick Ob- 
servatory Bulletin, 5, 15, 1908, 


of coronium and a strong continuous spectrum. In the less 
intense outer corona, the Fraunhofer lines are seen. In the 
inner corona, due to proximity to the sun's surface, are 
found temperatures approximating those in the photosphere. 
On account of minute pressures existing in the corona, the 
atoms may readily lose one, two, or more external electrons. 
The long free paths permitted by these small pressures 
cause conditions that cannot be approximated in our terres- 
trial laboratories. Unfortunately, due to the paucity of light 
and the small amount of time available for investigations, 
coronal wave-lengths are much less accurately determined 
than those of the nebulae. The source of the coronium 
lines will unquestionably be found among the elements in 
the top part of the periodic table. The " coronium " prob- 
lem will be more difficult than that of " nebulium." 

As the coronium emission constitutes a very small part of 
the total energy found in the corona, it may practically be 
left out of consideration in attempting to find a theory to 
explain the radiation of the corona. Any adequate theory 
must give solutions to the following problems: (i) What is 
the cause of the strong continuous spectrum of the inner 
corona, and ( 2 ) What is the explanation of the Fraunhofer 
spectrum in the outer corona? 

The first theory of the corona to be propounded was the 
one, so very popular at the time and thought to be capable 
of explaining away most of the astronomical problems 
the meteoric hypothesis. In virtue of this theory, the corona 
is nothing more nor less than the trails of myriads of meteors 
as they fall into the sun. Even at the time 1 a great au- 
thority on meteors, Newton of Yale, pointed out that the 
details observed in the corona were " inconsistent with any 
conceivable arrangement of meteoroids in the vicinity of the 
sun." The hypothesis was such an artificial one and had 
so little to commend itself that it is surprising that it found 
such a large place in astronomical literature but then we 
must not forget that very little was known concerning the 

Schaeberle's mechanical theory of the corona has much 

1 Nature, Sept. 30, 1866. 


more to recommend it but it too does not seem to have 
been able to bear the test of time, at least not in its original 
form. By virtue of this theory, the corona is caused by 
light emitted and reflected from streams of matter ejected 
from the sun by forces acting along lines normal to the sur- 
face of the sun. The forces are most active near the center 
of the sun-spot zones, and consequently, are confined almost 
wholly to the equatorial regions. Hence, as a result of this 
theory, the rays seen around the poles of the sun can have no 
existence, except that of streamers from the equatorial re- 
gions seen projected by perspective above and below the 
poles. In order that the force of ejection may be suffi- 
ciently great to overcome the attraction of gravitation, it 
was necessary to ascribe to the materials forming the longest 
rays initial velocities as large as 400 miles per second. While 
velocities of this size are not impossible on the sun, the truth 
of the matter is that no such motions have ever been discov- 
ered in the corona. Hence it is necessary to discard the 
theory or to modify it in some essential details. Moreover, 
" according as the observer is above, below, or in the plane 
of the sun's equator, the perspective overlapping and inter- 
lacing of the streamers cause the apparent variations in the 
type of the corona." This explanation might satisfy an 
annual variation (which is not known to exist), but fails to 
account for the change in form coincident with the eleven- 
year sun-spot period. 

There are, unquestionably, many forces acting on the co- 
ronal materials repelling the particles away from the sun 
in opposition to gravitation, and as these forces have been 
considered one by one, different coronal theories have been 
propounded. In 1885 the " electrical theory " was announced 
by Huggins. 1 The eclipse of 1889, of the minimum sun-spot 
type, having exhibited strong polar rays much resembling 
the lines of force about a magnet, Bigelow L> brought forward 
the " magnetic theory/' and Ebert 3 the " electro-magnetic 
theory." Bigelow's theory is very successful in explain- 

1 Proc. Royal Society, 39, 108, 1885. 

2 The Solar Corona discussed by Spherical Harmonics, 1889. 

3 Astronomy and Astrophysics, 12, 804, 1893. 


ing the details of the minimum type, but not so fortunate 
with the other forms of corona. The recent investigations 
at Mt. Wilson have shown that a magnetic field does exist 
around the sun, as demanded by Bigelow's theory. 

Great have been the claims of the exponents of the " ra- 
diation-pressure theory " not only for explaining the details 
of the corona, but also for furnishing a rational elucidation 
of why comets' tails always point away from the sun, and 
what causes the aurora borealis, zodiacal light, etc. That 
a ray of light exerts a pressure on any surface on which it 
impinges comes as a direct result of the Electro-magnetic 
Theory of Light published in 1873 by Clerk Maxwell 1 and, 
as was shown by Bartolr in 1876, can be deduced from 
the second law of thermodynamics. 

Since the light-pressure depends directly on the intensity 
of the sun's radiation, which decreases inversely as the 
square of the distance, as is also the case with gravity, the 
ratio of pressure to weight is therefore a constant independ- 
ent of the distance from the sun. The manner in which 
this ratio is found was first shown by Bessel 3 in 1836, who 
computed the magnitude of the repulsive force from the 
curvature of the tail of the comet in 1811. Bredichin, 4 
more recently, from measures of many comets' tails, has 
found them to be of four different types, in which the re- 
pulsive forces are respectively 18.5, 3.2, 2.0, and 1.5 times 
the attraction of gravity; the straight tail, according to his 
ideas, consisting of hydrogen, the plume-like tail of hydro- 
carbons, and the short stubby one of metallic vapors, chief 
among which are iron and sodium. The electrical force, on 
which Bredichin explains his repulsions, has been shown by 
Lebedew 5 not to have a sound physical basis. 

This objection cannot be raised to the principle of Ar- 
rhenius. That light actually exerts a pressure has been 
shown by Lebedew, 6 and Nichols and Hull 7 ; the latter, in- 

1 Electricity and Magnetism, 792. 
- // Nuovo Cimento, 15, 195, 1883. 

3 A. N., 13, 185, 1836. 

4 Annales de VObservatoire de Moscou, (2), i, 45, 1886. 
r> Astrophysical Journal, 16, 155, 1902. 

6 Annalen der Physik, (4), 6, 433, 1901; Astrophysical Journal, 15, 60, 1902. 

7 Physical Review, 13, 307, 1901; Astrophysical Journal, 15, 62, 1902. 


deed, have succeeded l in producing a laboratory comet's 
tail, although, as pointed out by them, other forces than 
light-pressure probably helped to give the repulsion. 

However, a rigid application of the theory of Maxwell 
is possible only when the body acted upon is large compared 
with the vibrations of light itself. When the body is of a 
size approximating the wave-length of light, Schwarzschild " 
has shown that the maximum value of the repulsive force is 
about twenty times the attraction of gravity. 

Undoubtedly, the radiation-pressure theory has been of 
the very greatest assistance in dealing with the corona, for 
the reason that it provides us with a knowledge of an addi- 
tional force acting in a direction in opposition to gravitation. 
Miller has published a series of excellent papers 3 enquiring 
into the question whether the coronal streamers exist in 
accordance with a modified Schaeberle mechanical theory, 
that their motions are produced by ejection, by the rotation 
of the sun, by the attraction of the sun and by the radiant 
pressure of the sun. For the purpose of the investigation, 
Miller examined the excellent series of photographs of the 
corona obtained by the Lick Observatory expeditions from 
1893 to 1918, inclusive, also plates secured by himself in 
1905 and 1918, and Lowell photographs taken in 1918. 
The conclusions drawn are that the force of repulsion is 
surprisingly large, being almost equal to the attraction of 
gravitation, and as a result, it is unnecessary to assume the 
very large velocities of Schaeberle's original theory. The 
facts accumulated seem to be in fairly satisfactory accord 
with the theory of Arrhenius expressed as follows: * " It 
is very probable that those drops, for which gravitation 
is just compensated by the pressure of radiation, will be the 
chief material of the inner corona. For drops of other sizes 
are selected out, the heavier ones by falling back to the 
sun, the lighter ones by being drawn away by the pressure 

1 Astro physical Journal, 17, 352, 1903. 

2 Sitzungsberichte der math.-phys. Classe der k. b. Akademie der Wissen- 
schaften zu Munchen, 31, 293, 1901. 

3 Astrophysical Journal, 27, 286, 1908; and 33, 303, 1911; and also Publ. A. 
S. P., 32, 207, 1920. 

4 Lick Observatory Bulletin 3 i, 152, 1902. 


of radiation, so that just those drops which, so to say, swim 
under the equal influence of gravitation and pressure of 
radiation will accumulate in the corona." 

By properly choosing his mathematical constants, it was 
possible for Miller to reproduce each and every coronal 
streamer so far discovered on eclipse photographs, even those 
streamers found near the sun's poles. Hence, the mechanical 
theory supplemented by that of radiation pressure seems to 
have solved many of the perplexing questions regarding the 
corona, but there are still many difficulties to be surmounted 
before we can feel ourselves on thoroughly sure ground. It is 
unquestionably necessary to take into consideration mag- 
netic and electric forces. Like radiation pressure and gravi- 
tation, these forces modify the other effects in a quantitative 
manner. 1 

That matter can exist in the finely divided state required 
by the theory of Arrhenius was shown 2 by " the eruption 
of Krakatoa, which drove the fine ashes up to an eleva- 
tion of 30 km ( 1 8 miles ) . The finest particles of these ashes 
were slowly carried by the winds to all parts of the earth, 
where they caused, during the following two years, the mag- 
nificent sunrises and sunsets which were spoken of as ' the 
red glows/ This glow was also observed in Europe after 
the eruption of Mount Pelee. The dust of Krakatoa further 
supplied the material for the so-called luminous clouds of 
the night/ which were seen in the years 1883 to 1892 float- 
ing at an elevation of 80 km (50 miles), and hence illumi- 
nated by the light of the sun long after sunset." 

If the temperature of the inner corona is approximately 
5000 C, it is quite difficult to see, as was pointed out by 
Abbot;" how matter can exist in the solid or liquid state 
or how the dust-particles of the inner corona can be " drops 
of liquid metal." 4 Another criticism of almost insurmount- 
able character is that voiced by Eddington. 5 Owing to the 
fact that conditions in close proximity of the sun cannot 

1 See Pringsheim, Physik der Sonne, 330, 1910. 

- Worlds in the Making, 7. 

:t Lick Observatory Bulletin, 5, 20, 1908. 

4 Ibid., 2, 1 88, 1904. 

5 Monthly Notices, R. A. S., 80, 723, 1920. 


be duplicated in the laboratory, we are ignorant of the true 
laws of radiation-pressure, which may have " encouraged 
quite exaggerated ideas of the possible effects of radiation- 
pressure.'' The " upper limit " to its power of supporting 
or driving out matter has been calculated by Eddington, 
and has been found to be equivalent to a pressure of 2 dynes 
per sq. cm. This can be likened to a wind of this strength, 
and the exact effect of any material will depend on its power 
of absorption of stopping the wind instead of letting it 
blow through. Allowing an ample margin for uncertainties 
of observation, Eddington calculates that the " pressures 
of radiation cannot carry a total weight of more than a milli- 
gram per sq. cm." Applied to the chromosphere, the den- 
sity is found to be of the approximate size of io~ 12 , a quantity 
which indeed appears so absurdly small that it seems to 
contradict all our ideas concerning prominences. The den- 
sity of the corona and in comets' tails on the same hypothesis 
is a thousand times smaller, or icr 1 '. 

Apparently therefore, the radiation-pressure theory is not 
entirely free from perplexities, but these difficulties may not 
be entirely insurmountable. Before further progress is made 
we must ascertain, by observational means, the size and 
the number of particles per unit volume forming the corona 
at different distances from the sun, and also the tempera- 
tures according to the Stefan and Wien-Planck formulas. 
The luminosity of the bright inner corona is caused by par- 
ticles which are heated to incandescence by solar radiation 
and which scatter sunlight, the particles being subject to 
gravitation, to radiation pressure and to electromagnetic 
forces largely unknown. Owing to their proximity to the sun, 
these particles cannot be in the solid or liquid condition 
and must therefore exist in the gaseous state. The mole- 
cules of the coronal gases strongly illuminated by sunlight 
probably act like the fine particles of a fog in scattering light. 
According to Fabry/ the part of the luminosity of the corona 
which gives the continuous spectrum may be due to this 
diffusion and not to reflection by small particles. As the 
result of some experiments on the diffusion of light by air, 

1 Observatory, 41, 211, 1918. 


Fabry estimated that a truly gaseous corona having a 
density only one-thousandth-millionth that of air at at- 
mospheric pressure would scatter sufficient sunlight to ac- 
count for the luminosity of the corona, and the polarization 
effects which have been observed. The explanation by 
Fabry has very much to commend it, but it is very much 
to be doubted whether it is based on sufficient experimen- 
tal evidence. Moreover, it is difficult to see why the light 
of the corona, if caused by molecular scattering, is white in 
color and not blue like the sky. 

Since the heavenly bodies, the sun, the stars, nebulae, 
etc., are bodies at high temperatures radiating energy, a 
knowledge of radiation laws is of vital importance to the 
solution of astronomical problems. Unfortunately until re- 
cently, little was known of these laws, but with the advent 
of the Bohr atom and Saha's theory of ionization, but spe- 
cially on account of the enormous activity in the past decade 
in investigating problems of radiation, we have already made 
many discoveries and many more are to follow. 

Saha's theory of ionization (see Chapter XVII) has al- 
ready been very successful in interpreting the importance of 
enhanced lines in the flash spectrum and in furnishing an 
adequate explanation of the differences between the Fraun- 
hofer spectrum, the sun-spot spectrum and the chromo- 
spheric spectrum. Cannot this same theory be expanded 
so as to furnish an adequate explanation of the cause of radia- 
tion in the corona? 

It has been very difficult to understand the process where- 
by the coronal particles are enabled to emit light at the very 
great distances of ten million miles from the surface of the 
sun, under conditions of very low pressure and very long 
mean free path. But the corona is an appendage of the sun, 
and somehow or other (we have never realized just how) it 
may be possible for the corona to borrow some of its radiation 
from the sun. But how about a body like the Orion nebula? 
It gives a bright-line spectrum. Are astronomers to keep on 
saying, as they have for a number of years, that the Orion 
nebula emits light because the kinetic energy causes the 
particles to be heated and that the luminosity is wholly the 


result of heat? It is excessively difficult to imagine a body 
of such vast dimensions heated to luminosity but yet sur- 
rounded by the intense cold of inter-stellar space. Ap- 
parently some cause other than that of temperature must be 

In studying the problem of galactic nebulae, Hubble 1 
comes to the conclusion that their luminosity is derived from 
the radiation from stars in the vicinity. If this were a 
simple case of reflected starlight, the spectra of the nebula 
would agree exactly with the spectra of the associated stars ; 
but such is found not to be always the case. Hubble re- 
marks, " It is doubtful whether or not a mass of diffuse 
nebulosity isolated in space and with no stars involved could 
hold together and at the same time shine by light generated 
by collisions of molecules. At temperatures corresponding 
to intensity-distribution or width of lines in nebular spectra, 
the average speeds of the molecules would be so high com- 
pared with the velocities of escape that the nebulosities 
would probably dissipate rapidly. On the other hand, if 
molecular speeds were sufficiently small to admit of cohesion 
in the mass, the nebulosity would probably be too cold to 
radiate light. This argument suggests that diffuse nebu- 
losity is not intrinsically luminous, but is rendered so by 
external causes." 

If it is possible to discover the mechanism whereby the 
nebulae are rendered luminous, the same mechanism will 
probably explain the cause of the coronal radiation. The 
fact must not be overlooked that each and every star has 
a corona surrounding it, but these coronas can never be 
rendered visible to astronomers since the total light of each 
corona must be comparatively feeble compared with the 
luminosity of the sun which it surrounds. Milne 2 has shown 
that the light of the corona can have no sensible effect in 
modifying the spectrum of the sun. 

The mechanism for explaining the radiation of corona 
and nebulae unquestionably is found in the electron. The 
sun and the stars are vast radiating spheres at very high 

1 Mt. Wilson Contributions, No. 241, 1922. 

2 Phil. Trans. Roy. Soc'y, A, 223, 201, 1922. 


temperatures, and from each of these suns, billions of elec- 
trons are shot off every second of time. The number of the 
electrons emitted depends on the intensity of the radiation 
while the energy of the electrons depends primarily on the 
temperature. The energy of the electrons will carry them 
to the greatest distances from the radiating body in the case 
of the very hottest stars. 

Schwarzschild proposes a theory with attractive possi- 
bilities, namely, that the corona consists of " electron-gas," 
that is, a gas of very long mean free path which is capable 
of reflecting and polarizing light. But as Schwarzschild has 
pointed out, the electrons, each of which carries a unit 
negative electric charge, must require an equally large posi- 
tive charge on the surface of the sun; this causing such a 
strong electrical field that even the fastest moving electrons 
would be stopped at distances less than a millimeter. In 
consequence, the corona could have but little extension. 
Quite independently, Mitchell (2d ed., p. 356) comes to the 
conclusion that the radiation of the outer corona is caused 
by electrons. Ludendorff attempts to get rid of the great 
electrostatic charges required by the electron-gas theory by 
assuming that atoms carrying positive charges must be 
mixed with the free electrons. 

In a series of excellent articles, 1 Wilhelm Anderson reviews 
the various coronal theories. He accepts the electron- 
gas theory. After making certain plausible assumptions con- 
cerning conditions that exist in the corona, he calculates, on 
the basis of convective equilibrium, the effective molecular 
weight of the coronal material, and he finds a good agree- 
ment with the atomic weight of the electron. The thermal 
radiation of the inner corona is naturally greatest near the 
solar surface, so, likewise, is the intensity of the light of the 
photosphere reflected by the electrons. Anderson takes 
the three laws of intensities of radiation in the corona, the 
inverse sixth, fourth and square, and he finds a good agree- 
ment between the two latter at a distance of 0.28 solar radii. 
On the assumption that at this distance the diffuse reflected 
light has 38 per cent of the intensity of the total illumina- 

1 Zeits. jur Physik, (4), 33, 1925. 


tion, he calculates the results from theory and finds a good 
agreement with the fourth-power law close to the sun, but 
at greater distances out a better agreement with the inverse- 
square law. 

It therefore seems to be quite probable that the coronal 
radiation takes its origin in the electron, the energy coming 
from two separate causes: (i) from thermal radiation re- 
sulting from collisions, and (2) by reflection and scattering 
of the photospheric light. The first effect is visible only in 
the inner corona, while the second manifests itself both in 
the inner and the outer corona. 

In quite similar manner, Zanstra 1 has made an applica- 
tion of the quantum theory to explain the luminosity of 
diffuse nebulae. The nebula appears to have little radiant 
energy of its own, but borrows or takes energy from the star 
(or stars), the electron being the medium for the transfer 
of energy. Bowen ~ explains the spectrum of nebulium as 
the result of the long mean free paths of the electrons, the 
jumps being " forbidden " under laboratory conditions. 

Hence, it seems to be definitely proven that both the solar 
corona and the nebulae are rendered visible as the result of 
the action of the electrons. Whatever the exact mechanism 
may be, both the number of electrons and their intensities of 
emission depend on the temperatures of the stars. Accord- 
ing to the investigations of Eddington and Milne, radiation 
pressure greatly assists in the discharge of electrons. 

It is unnecessary to assume with Schwarzschild that the 
corona consists almost exclusively of electron-gas. Zanstra 
made no such assumption in explaining the radiation in nebu- 
lae. In the inner corona is found the emission spectrum due 
to coronium. This radiation can be caused only by the colli- 
sions of electrons. But colliding with what? Evidently 
with protons and with atoms which have lost one, two, or 
more external electrons. Above the i4,ooo-km level reached 
by the highest lines of the chromosphere, there are unques- 
tionably other ions reaching far greater heights but whose 
radiations give no light in the visible spectrum but are found 

1 As trophy steal Journal, 65, 50, 1927. 

2 Ibid., 67, i, 1928. 


in the ultra-violet beyond the reach of spectroscopic investi- 
gations. Coronium atoms have been detected to distances 
twenty times greater than the highest in the chromosphere. 
Hence, in the inner corona it is necessary to assume ionized 
atoms and protons. These are most numerous near the solar 
surface. They thin out rapidly in the inner corona and few 
of them are found in the outer corona. In the outer corona 
there can be relatively few ionized atoms, and hence the 
outer corona is visible mainly as the result of scattering of 
the photospheric light by the electron. 

Regarding the shape of the corona, we must recall that 
near sun-spot minimum the spots in the dying cycle are near 
the solar equator while those awakening into life are at 
higher latitudes. Hence when the sun is quiescent, the effect 
will be that of very long streamers going out in straight lines, 
the maximum lengths being attained not at the equator but 
at the higher latitudes of the awakening spot zone, a fact 
already noticed by Campbell. A spot on the sun, or an 
active prominence, may be a local center of activity on the 
sun, with the result that coronal streamers or hoods sur- 
rounding the prominences may result. Since the field 
strength in spots is far greater than that of the sun as a whole, 
the comparative inactivity of the sun at its poles is exhibited 
by the short polar brushes. Even before passing the epoch 
of minimum of spots, the sun takes on new energy, and hence 
electrons are discharged with increasing strength in regions 
not necessarily limited to the solar equator. Consequently, 
the corona takes on first a rectangular shape, and then a con- 
tour more and more circular as the time of sun-spot maxi- 
mum is approached. Owing to the greater average vigor of 
electronic discharge at sun-spot maximum, there should be a 
greater intensity of the emission lines of coronium than at 
minimum of spots; and this is found actually to be the case. 
After passing through the minimum of spots, the awakened 
solar activity shows itself in three different portions of the 
sun: (i) In the photosphere, the increased radiation causes 
spots to appear. (2) In the chromosphere, the increased ra- 
diation carries the elements of medium height to greater aver- 
age elevations. (3) In the corona, the increased radiation 


causes an increase in strength of the emission lines of coro- 
nium and also makes the corona lose the shape associated 
with minimum of spots, of long equatorial extensions and 
short polar brushes. 

According to the recent work of many investigators, there 
is some doubt existing as to whether the atmospheres of stars 
are in thermodynamic equilibrium. If they are not, then 
the result will be that the theoretical conclusions now ( 1932 ) 
believed to be true must probably be modified in essential 
details. Furthermore, as conditions existing in the sun can- 
not be approximated in laboratory experiments, little is 
actually known of the exact physical laws involving the dis- 
charge of electrons from the sun or those underlying radia- 
tion pressure at the solar surface. It seems almost useless 
to test any theory from laboratory experience. 

It is now more than fifty years since the first attempt was 
made by Huggins to photograph the corona in full sunshine. 
The authority of his name, great in the annals of spectros- 
copy, gave a degree of plausibility to the problem. The 
task to be overcome is to separate the light of the corona 
from the strong illumination of the sky. The chief names 
connected with this work are those of Hale, Ricco, Des- 
landres, Wood and Hansky. It was natural that the meth- 
ods so successful in photographing the prominences should 
be first tried, and, in order that the atmospheric glare might 
be reduced as much as possible, the observations were made 
from mountain tops, Pike's Peak and Mt. Etna being oc- 
cupied for this purpose. No success being secured, a series 
of attempts were made by heat-measuring instruments like 
bolometers or thermopiles. Photographic methods, of using 
color filters and plates sensitive to different parts of the spec- 
trum, have been thoroughly tested. Since the Great War, 
when such noted success was attained in airplane photog- 
raphy by using plates sensitive to the deep red, attempts were 
revived on the corona. Lindemann had found it possible to 
photograph stars of the first magnitude near the sun. Each 
and every one of the plans, however, at times carried out with 
great skill and ingenuity, have always ended in the same 
manner failure to photograph the corona in full daylight. 


The observations by Abbot in 1908 and those by Kunz and 
Stebbins in 1918 have shown the cause of the failures, namely , 
the intrinsic feebleness of the corona. Even in the brightest 
parts, the inner corona is no brighter than the surface of the 
full moon, which has a brilliancy six-hundred-thousand times 
less than the sun. The corona is about equal to the intensity 
of the illuminated sky at eight or ten degrees distant from 
the sun, but close to the sun's edge the light of our central 
luminary is so overpowering that it appears indeed well- 
nigh impossible to photograph the corona in full daylight. 

In view of the above conclusions, it is interesting to read 
of the recent work 1 of Lyot of the Meudon Observatory. At 
the summit of the Pic du Midi, at an altitude of 2800 meters, 
he used a telescope stopped down, and with a disk in the 
focal plane extending 30" beyond the sun's edge. Observ- 
ing directly with an eyepiece, he saw rose-colored promi- 
nences. Setting the slit of a spectrograph close to the limb, 
he was able to see and to photograph the 5303 and 6374 
coronal lines, superposed on the Fraunhofer spectrum. Po- 
lariscopic observations were made and direct photographs 
were obtained which resembled in appearance the inner co- 
rona. It is surprising to find success where others had failed. 
The work is beset with such great observational difficulties 
that future researches will be followed with the greatest of 

The past decades have been golden years for the progress 
of astronomy, particularly on account of the attack on 
atomic structure by the astronomer, physicist and chemist, 
combined. The importance of the electron has thus been 
recognized. The theory of ionization which has already been 
so successful in furnishing an explanation for many of the 
difficulties connected with the flash spectrum, the chromo- 
sphere and sun-spots is but a branch of a larger theory of 
photo-electricity dealing with the production of light by the 
passage of electricity through gases. Photo-electric action 
involves both ionization and radiation. When an electron 
strikes an atom and a transfer of energy takes place, there 
may be complete ionization, as shown by the production of 

1 Comptes Rendus, 191, 834, 1930; and 193, 1169, 1931. 


positive and negative ions, or there may be partial ionization, 
that is, a disturbance of an atom which is not detectable as 
ionization but is shown by the production of radiation. Both 
radiation and ionization are caused by the action of electrons 
in their bombardment of atoms. The fundamental basis as- 
sumed for Saha's remarkable theory is that the electrons obey 
the same laws as gases, or, in other words, that they have the 
same physical properties as the atoms. 

Moreover, recent work in astrophysics has demonstrated 
conclusively that the source of energy is found in the electron, 
that even mass itself has its origin in the electron. Edding- 
ton has shown that the stars are all slowly losing mass, for 
the reason that radiation is energy, the dissipation of energy 
means the discharge of electrons, which is synonymous with 
saying a diminution of mass. Consequently the sun and stars 
are continually losing mass by the discharge of electrons. 
Hence it is inevitable, as a result of this theory, that unless 
energy is being created possibly by cosmic rays the end 
of the universe may be foreseen. 

It is unfortunate that the corona can be investigated only 
at the rare occasions of total eclipses and that an individual, 
no matter what his enthusiasm or skill, will have during his 
whole lifetime less than one hour within which to make all of 
his observations of the corona. In the immediate future, 
total eclipse expeditions mean long trips from home to loca- 
tions where there is little reason to expect good atmospheric 
conditions; hence progress in knowledge concerning the co- 
rona promises to be very slow. 



EINSTEIN'S theory of gravitation has been justly re- 
garded as the greatest triumph of mathematical 
reasoning that has taken place since the time of 
Newton. It is safe to say that no scientific achievement of 
recent years has aroused so much popular interest and en- 
thusiasm as that evoked by the verification of the Einstein 
prediction from observations made at the total eclipse of 
May 29, 1919. 

The first step in any scientific investigation is to get at 
the facts, derived from observation usually by a series of 
measurements. Great precision, patience and care are 
necessary to enable the observer to record the true facts 
devoid of any inferences or illusions of the mind. Lord 
Kelvin has said that, " Nearly all of the greatest discoveries 
of science have been but the rewards of accurate measure- 
ment and patient, long-continued labor in the minute sift- 
ing of numerical results." After the observations'. have been 
secured, they are classified and analyzed and tested to see 
whether they will conform to some known law. The laws 
of nature differ greatly from human or civil laws since the 
former must be universally true and must apply under all 
conditions. If a law of nature is found to be deficient, even 
in some minor detail, it must be revised to satisfy the con- 
ditions, or, if this is impossible, then the law must be dis- 

The grandest and best known natural law is that of 
gravitation discovered by the great Newton. Its importance 
lies primarily in the fact that it applies to all bodies, it is 
universal. In the two and a half centuries that have elapsed 
since the falling of the apple, the law of gravitation has 
experienced one grand triumph after another, with the in- 



evitable result that we have come to look upon this law as 
practically infallible. Other laws may come and go and be 
revised into changed form, and go on again to be again and 
again revised, but the law of gravitation has stood firm 
without change, with the enunciation of its principle in the 
same form as when handed to us by Newton. Every body 
in the universe attracts every other body in the universe 
with a force that is proportional to the product of the masses 
and inversely to the square of the distances apart. 

And now we are told that the law of gravitation must 
be discarded, that space is curved and that we live in a 
world of four dimensions. The mind reels, and refuses 
to be convinced. Common sense tells us, we say, and 
the experience of the ages has proven, that space has but 
three dimensions, length, breadth and thickness, so why 
say such an apparently foolish thing that time, which we 
know has nothing to do with space, must be considered as 
a fourth dimension? The average man of intelligence im- 
mediately calls to mind some of the properties of the mathe- 
matician's space of four dimensions, a man enclosed in 
a steel-proof vault, if living in a four-dimensional space, 
could get outside of the vault without passing through any 
of the walls. One smiles incredulously, and thinks of a re- 
mark by Bertrand Russell that " mathematics may be de- 
fined as the subject in which we never know what we are 
talking about, nor whether what we are saying is true." 

Is the theory of relativity of Einstein an unreality in the 
brain of the mathematician, or must we accept it as true 
and believe in a four-dimensional space? If the Einstein 
theory is accepted, then as an important consequence the 
law of gravitation must be revised, there is no middle course. 
The new theory states that there is no " force " of gravita- 
tion, for this " force " is but one of the inherent properties 
of four-dimensional space. To decide which of the two 
conflicting hypotheses best represents the law of gravitation, 
Newton's or Einstein's, we shall have to pass in review the 
salient facts. Science desires the simplest explanation. 
Einstein's own conception of the law of gravitation, as ex- 
pressed in the New York Times is, " Please imagine the 


earth removed, and in its place suspended a box as big as 
a room or a whole house, and inside a man naturally float- 
ing in the center, there being no force whatever pulling him. 
Imagine further, the box being, by a rope or other con- 
trivance, suddenly jerked to one side, which is scientifically 
termed ' difform motion,' as opposed to ' uniform motion.' 
The person would then naturally reach the bottom on the 
opposite side. The result would consequently be the same 
as if he obeyed Newton's law of gravitation, while, in fact, 
there is no gravitation exerted whatever, which proves that 
difform motion will in every case produce the same effects 
as gravitation. 

" I have applied this new idea to every kind of difform 
motion and have thus developed mathematical formulas 
which I am convinced give more precise results than those 
based on Newton's theory. Newton's formulas, however, 
are such close approximations that it was difficult to find 
by observation any obvious disagreement with experience." 
Does this theory of Einstein concerning a man falling from 
the top of a step-ladder to the floor of a room really furnish 
a simpler explanation than the classical law of falling bodies 
of Newton? If it does, then a very radical change in our 
modes of thought must take place. 

History shows that the present is not the first time that 
the perfect science of astronomy has been forced to change 
its point of view. In the early days of civilization it was 
believed that the earth was flat, and this notion is still 
not quite eradicated from human thought. Later was 
evolved the theory that the earth was the center of the 
universe and about it revolved the sun, moon, planets and 
stars. The mechanism required by the Ptolemaic system 
to explain the motions of the heavenly bodies by cycles and 
epicycles was such a cumbrous one, that it led to a remark 
by a former king of Spain that if he had been present at 
Creation he could have furnished the Creator with some 
very valuable suggestions. The work of Copernicus, Tycho, 
Kepler, Galileo and Newton revolutionized scientific thought, 
and as a result the earth was displaced from her important 
position as the center of the universe. Although the geo- 


centric point of view is still retained in some of our scien- 
tific explanations, we realize that the sun is the center of 
our system, with the earth occupying no more exalted a 
position than that of being but one of the planets. The 
Newtonian mechanism reduced celestial motions to the 
greatest simplicity. In obedience to the law of gravitation, 
every planet describes a perfect ellipse about the sun as 
focus, and these elliptical orbits would repeat themselves 
indefinitely were it not for the gravitational forces arising 
from the other planets. After allowing for all of the 
planetary disturbances, or " perturbations/' Newcomb found 
that the orbit of the innermost planet Mercury was rotating 
in its own plane at the rate of 42 seconds of arc per century 
in excess of that required by the traditional theory. Vari- 
ous attempts were made to explain away this discrepancy 
in the motion of Mercury by assuming a planet or planets 
inside the orbit of Mercury, or outside of its orbit, by postu- 
lating a belt of matter in a flattened disk or series of ellipses 
surrounding the sun, or by presupposing that the sun has an 
oblateness of shape and distribution of mass different from 
that usually taken for granted but all to no avail. In 
every case, the assumption of mass required to produce the 
observed effect on Mercury would have caused disturbances 
not observed in the other planets. The solution of the 
problem came only with the theory of relativity which fur- 
nished a motion of the perihelion of Mercury of 43 " per 
century, in almost exact agreement with the observed dis- 
crepancy. And so, when in addition, the observations of the 
1919 total eclipse gave deflections in accordance with the 
Einstein prediction, the confirmation of the theory of rela- 
tivity seemed almost complete. 

Newton's law, however, is not invalidated by the latest 
discoveries but rather is supplemented by them. It is pos- 
sible in very few instances, and as the result only of the 
most refined measurements, to be able to detect the differ- 
ence in observed effect between the Einstein and the New- 
tonian laws. In addition to his great work on gravitation, 
Newton proposed an erroneous theory of light. According 
to this theory, light consists of minute corpuscles expelled 

Upper Pair: Exposures 10 sec. (left) and 20 sec. (right). 
Lower pair: Through Nicol prism turned 90 between exposures, each of 5 sec. 


from a luminous source with high velocities. In the estima- 
tion of Kayser, 1 " Newton's errors impeded science even to 
the middle of the nineteenth century." It was not until 
the observations of Hooke, Huyghens, Young and Fresnel 
showed the corpuscular theory to be untenable that the 
wave theory was finally established. But how do the waves 
of light travel from the sun to the earth? It was necessary 
for the physicist to invent some material medium for their 
propagation, and hence the existence of an imponderable 
and universal " ether " was devised. With the discovery 
of the laws of electromagnetism, it was shown by Maxwell 
that electromagnetic disturbances also were propagated with 
the speed of light, and the conclusion was drawn that light 
itself must be electromagnetic in character. Additional 
functions were imposed by physicists upon the universal 
medium with the result that Maxwell and Kelvin were able 
to make estimates of the density and rigidity of the ether 
with almost as much assurance as if they had been dealing 
with perceptible matter. (More recent researches have 
shown that not only is light electrical in nature but so also 
is mass and energy.) 

Lis;ht and all electromagnetic manifestations were believed 
to be actions taking place in the all-pervading ether. Light 
from the distant stars reaches us in the form of wave mo- 
tions in the ether, and it has been necessary to assume that 
the sea of ether between us and the distant stars is un- 
broken. Moreover, in order to account for astronomical 
aberration and other phenomena, it was proven beyond doubt 
that the ether must necessarily be at rest, that it pervades 
everything but that it is not dragged about by ponderable 
material. Since motions of all kinds are relative in their 
nature, it was of the utmost importance to theories of physics 
that a state of absolute rest should exist in nature and 
this appeared to be found in the condition of the ether. 
Accordingly, attempts were made to detect the drift of the 
earth through the motionless ether. In 1887, the first ex- 
perimental determination was carried out by Michelson. 
The same experiment was later repeated by Michelson and 

1 Handbuch der Spectroscopie, i, 5, 1900. 


Morley, by Morley and Miller, and quite recently by Miller 
alone. This experiment has been described so frequently 
that it will be unnecessary here to go into details. The 
velocity of light was known to be, in round numbers, 186,000 
miles per second, and this was supposed to represent the 
velocity of the wave motion through the ether. The velocity 
of the earth in its orbit is 18 miles per second, and hence if 
the ether is stationary with respect to the sun, the drift of 
the earth through the ether should be at this speed. Ac- 
cordingly, if a beam of light reaches the terrestrial observer 
in the direction of the earth's motion, it will be traveling 
with respect to and overtaking the earth at a speed of 
185,982 miles per second, while if the beam of light comes 
in the opposite direction its velocity with respect to the 
observer will be 186,018 miles per second. Although it is 
impossible to measure the velocity of light along a single 
straight course with sufficient precision to compare the two 
velocities above, nevertheless these velocities can be com- 
pared by a differential method. This was accomplished by 
the Michelson-Morley experiment. This consisted in di- 
viding a beam of light by means of an unsilvered mirror, 
allowing each half of the beam to travel the same distance 
in directions perpendicular to each other, and then reflecting 
the two beams back to the same point at which they were 
originally separated, and then recombining them. It is evi- 
dent that on account of the motion of the earth, the path 
through the ether of the half beam which is traveling paral- 
lel to the direction of the earth's orbital motion must be 
slightly longer than the other half of the beam. Conse- 
quently, when the two halves of the beam are reunited, the 
waves of light should be a little out of phase. By turning 
the apparatus through a right angle, a check is obtained on 
the equality of the distances traversed by the two beams. 
The precision of the experiment was so great that it should 
have been possible to detect the drift of the earth through 
the ether equal to one-tenth of its orbital velocity about 
the sun. But no effect whatever was detected and there 
was no indication that the earth was moving at all through 
the ether. 


The most obvious explanation of the failure of the 
Michelson-Morley experiment to detect the drift of the 
earth is that the ether is stationary with respect to the earth, 
or is carried along by it. But if we admit this possibility, 
we are in a quandary to explain the aberration of light from 
the stars. Everyone is familiar with the fact that on a rainy 
day when there is no wind the rain falls vertically, and to 
keep dry, when standing still, it is necessary to hold an um- 
brella over one's head. If the holder of the umbrella starts to 
walk he tilts the umbrella forward in the direction of his mo- 
tion, for experience has told him that only by so doing can he 
keep dry. The faster he walks the more theumbrella is tilted 
forward. This is a simple case of relative motion. So when 
a telescope is pointed in the direction of a star, the telescope 
tube must be tilted forward by an appreciable angle so that 
the rays from the star may pass down the tube to the eye 
of the observer. The angle through which the telescope has 
to be inclined is known as the angle of aberration. This 
angle was discovered by Bradley in 1728. Its value is 
2o"47. On account of the great precision of modern astro- 
nomical measurements this angle is considered large. 

But how explain the Michelson-Morley paradox that the 
earth must be drifting through the ether while at the same 
time the effect of this drift has never been detected? Ap- 
parently there is no way out of the dilemma but that our 
ideas must in some manner be revised. In 1893, a satis- 
factory explanation was offered by Fitzgerald, and in 1895 
independently by Lorentz, that the negative result of this 
experiment becomes quite intelligible if it is assumed that 
when a body is in motion through the ether its dimensions 
in the direction of motion become slightly shorter than when 
it is at rest. This would indeed seem to be a strange and 
a highly arbitrary hypothesis if unsupported by other scien- 
tific evidence. The work of Lorentz and Larmor on the 
electromagnetic theory showed that it was necessary to as- 
sume that a body when in motion does actually contract 
by just the amount demanded by Fitzgerald's explanation. 
If this hypothesis is true, then as a result we are inevitably 
forced to the conclusion that time also must be measured in 


a different manner for an observer at rest and for one in 
motion. The Fitzgerald contraction is not merely an 
idle arbitrary speculation, but after repeated experiments 
it has been shown to be true and mathematically exact in 
the well-known laws of electromagnetic forces. 

The amount of contraction depends on the velocities, 
and in the majority of the cases of our experience, or 
for average velocities, is excessively minute. In the case 
of the earth the contraction is only one part in 200,000,000, 
which corresponds to a contraction of 2^ inches in the 
earth's diameter in the direction* of its motion. The 
Michelson-Morley experiment therefore failed to detect 
the motion through the ether for the simple reason that 
the dimensions of the apparatus were actually automatically 
contracted by an amount just sufficient to compensate for 
the effect sought. Other ingenious experiments, electrical 
and optical, were tried for the purpose of detecting the drift 
through the ether, but always with the same result. There 
thus appears to be a " conspiracy " among the various 
agencies at work to prevent man from measuring his motion 
through space. 

In 1905, Einstein published his restricted Principle of 
Relativity that " it is of necessity impossible to determine 
absolute motion by any experiment whatever." According 
to Einstein, all motion is relative, and no experiment can 
be possibly devised so as to decide which of two systems 
is at rest and which in motion. It is therefore impossible 
by any experiment to detect uniform motion with respect 
to the ether. Hence the assumption of the ether, brought 
into physical science for no other purpose than to explain 
the propagation of light, becomes entirely unnecessary. If 
the ether has no position whatever in space, the statement 
that it exists has no meaning. Hence, if the ether is retained 
by the physicist he must assign a new set of properties. 
A second consequence is that the velocity of light with 
respect to two observers moving relatively to each other 
must always be the same, no matter in what direction the 
light is traveling, as the velocity must be equivalent to 
that when determined by an observer at rest, of 186,000 


miles per second. These radical changes can exist only 
under the condition that certain relations, known as rela- 
tivity transformations, exist between the space and time 
measurements made by the two observers. 

The principle of relativity thus entails changes in our 
mode of thought of the most revolutionary kind, and our 
preconceived notions in consequence are turned topsy-turvy. 
Let us see a few examples. If an observer is traveling at 
the rate of 161,000 miles a second in a vertical direction his 
arm, 30 inches long, is of its natural length when extended 
horizontally, but is contracted to 15 inches when hanging 
by his side or raised vertically above his head. This you say 
is foolishness. Well, take a yardstick and measure the 
length of the arm. Horizontally the arm measures 30 inches 
on the scale, and vertically exactly the same 30 inches by the 
scale. Yes, but when you hold the scale vertically it too 
has been contracted and each inch on the scale is in reality 
only half an inch, so that the arm measures 30 half -inches or 
15 inches. "Yes," you will say, "but whoever heard of 
anyone moving with such a colossal speed as 161,000 miles 
per second? " Why not? It is impossible to measure ab- 
solute velocities in space, and for all we know to the con- 
trary we may be speeding along at this terrific rate. The ft 
particles, referred to in Chapter XIV, move with velocities 
of 100,000 miles per second. 

A conclusion perhaps even stranger than the foregoing 
must be drawn from the principle of relativity. Suppose 
observer A is moving through space at a speed of 161,000 
miles per second relative to B, and allow each observer to 
hold an identical yardstick. On account of his fast flight, A's 
stick is contracted to half of the length it has when at rest, 
and hence B's stick appears to be twice the length of his 
own. But since motion is purely relative in character, B 
must be moving with respect to A with the speed of 161,000 
miles per second, and hence B's yardstick is contracted; and 
it appears to be twice as long, half as long, or equal to the 
length of an identical stick, depending on which stick is the 
moving one or whether they both are at rest. 

If the speed of the observer is increased beyond 161,000 


miles per second , the contraction becomes greater and still 
greater; and at the velocity of light all lengths dwindle to 
zero and vanish. The observer changes to an object of two 
dimensions, length and breadth, but without thickness, but 
he himself is blissfully unconscious of his sorry plight. Dis- 
tance is annihilated and so also is time. If traveling at the 
velocity of light, the observer could go to the most distant 
star without becoming a day or even a second older, for the 
reason that no time had elapsed. These paradoxes make 
the theory of relativity very difficult of comprehension to 
anyone who is not a mathematician and they seem moreover 
to controvert our " common-sense " view of things. 

Everyone will agree that if one measuring rod identical 
in length with another can be twice as long, half as long or 
equal in length to the other, there must be something wrong 
somewhere with the method of measurement. Einstein in- 
quired into our methods of measuring length and time and 
pointed out the lack of definiteness in the concepts of space 
and time as ordinarily used. What do we do when we meas- 
ure the length of an object like a book? We take a foot-rule 
or meter stick and determine its " true " length. We 
measure the size of the book, however, only by comparing it 
with a scale whose length we suppose to be known. But 
what do we actually know of the length of the foot-rule? 
We take it for granted that it has been compared with some 
standard foot-rule or yardstick. We know the lengths of 
our measuring rods only by comparing their lengths with 
other rods, and so on and on through an almost endless series 
of comparisons, until we come to the Standard Yard pre- 
served in London, or the Standard Meter kept in Paris. 
The latter is a rod of definite shape, of a certain particular 
metallic composition, and at a given temperature, the length 
between two specified lines is exactly a meter. The length 
of the meter is approximately one ten-millionth part of the 
earth's quadrant. What would happen if the earth and all 
it contains were suddenly contracted to half its present 
dimensions? All objects including ourselves and all measur- 
ing appliances would dwindle to half size. We would never 
have any idea that the earth had contracted if we kept our 


attention on terrestrial objects. But lo, and behold, the 
sun has suddenly doubled its size! If the sun and the ex- 
ternal universe could have contracted at the same time as 
the earth and the observer, we could never have become 
aware of the difference. The " true " length of a book or 
a rod thus depends on the definitions and postulates which 
we adopt. Manifestly there is no such thing as " absolute " 
space, everything is merely relative to the observer. 

You will insist, however, that if there is no " absolute " 
space, there must certainly be " absolute " time, for time 
seems to go on quite independently of any observer. Each 
of us unconsciously feels that when we die, the world will 
carry on just about as before. Time and space thus seem 
on a very different basis. But how do we measure time? 
Ordinarily by a clock. We assume that the pendulum swings 
in a perfectly uniform manner and that the clock hands move 
equal distances in equal intervals of time. But how prove 
our assumption? We are forced for our definitions of time 
intervals to go back to the earth, and to assume that the 
earth is a colossal top set spinning on its axis, and its mass 
is so great that like a gigantic fly-wheel its motion must cer- 
tainly be uniform. The assumption seems sound and all 
of the accurate measurements of astronomy are based on it. 
What would happen to the vaunted predictions of as- 
tronomy if the earth changed its speed of rotation? In fact, 
it was shown in Chapter IV that some suspicion has lately 
been cast on the uniformity of rotation of the earth, and 
our time-keeper may after all be a faulty one. Apparently 
therefore, we seem forced to the conclusion, since time- 
measurements depend on measures of distances, that there 
is no absolute time any more than there is absolute space. 
The determination of both space and time consist essentially 
in the measurement of certain lengths, and such measures 
have no meaning except in relation to ourselves and our 
every-day experiences. 

Let us return again to the measurements of the book. We 
recognize the object because we have seen other books. We 
glance inside the pages, and find a treatise on higher mathe- 
matics or some ebullition of modern fiction, both unreal and 


devoid of meaning to our feeble intellect. We likewise 
recognize the foot-rule and we accept some one's word for 
it that this rule has been compared with other rods, etc., etc., 
and that we know its length. Applying the measuring rule 
to the book, and spending a few seconds of time in the opera- 
tion, we find a length of eight inches. A Martian dropped 
to the earth would not recognize a book, for in the advanced 
civilization that is supposed to exist on the ruddy planet, 
books do not exist. Neither would he recognize the foot- 
rule, and certainly the marks that designate the inches would 
be utterly unintelligible. The simple process of measuring 
the length of a book would be completely impossible to the 
gigantic brain of the Martian. He has not been taught 
from his childhood up to measure with our particular rods. 

The measurement of a length, simple though it seems, is 
after all a very complicated process. We see the book, and 
note that it occupies a definite position on a table in a room 
furnished as a library. Having two eyes, we have learned 
unconsciously to estimate the distance to the book, and so 
we telegraph the thought to our brain that we must reach 
forth the hand and pick up the book. (It is not necessary 
to trace further the train of thought.) But it is easy to see 
that in measuring the length of the book we are putting into 
practice the training of hundreds and even thousands of 
years, acquired at times with great difficulty by ourselves 
and our ancestors. 

Knowledge thus comes to us only as a series of experiences 
or events. Each experience is perhaps an instantaneous 
photograph on the brain. But every photograph when taken 
requires a definite exposure-time, sometimes longer, some- 
times shorter in length, depending on the intensity of the 
light and the speed of the photographic plate. And so 
there are brains which work slowly and those which grasp 
more quickly. We see therefore that throughout all of our 
existence we have tacitly assumed, without realizing it, that 
measures of space are not at all independent of time, and 
that durations of time are not devoid of their inherent refer- 
ence to space. There are no durationless events, and each 
and every one of our experiences involves a consciousness 


of both space and time. Thus neither space nor time is 
absolute, and every observation is relative to the individual 
observer. But each observer being confined to the earth, 
by common consent we have come to the conclusion that 
Truth is attained only when our observations agree with 
that of the average individual and that we are in error when 
we differ from the general average. 

Throughout the whole of our scientific consciousness we 
have assumed that in agreeing with the average observer 
we have thereby derived the " true " length or have de- 
termined the " true " duration of time. The great triumph 
of Einstein consisted in making clear that the " true " 
measures of each observer, and hence of the average ob- 
server, are only relative in their nature and not absolute in 
value. Consequently, all of the observations with which 
we are familiar are relative and depend only on the indi- 
vidual as represented by the average observer. The method 
of advance was manifestly to divorce the series of events 
from the observer, and look upon them from the point of 
view of the objects themselves. This was Einstein's method. 
He assigned space and time solely to the observer, but gave 
to nature an unfamiliar combination of space and time of 
four dimensions. However, this is not the same four- 
dimensional space that has lately been much discussed by 
the mathematician, a space of length, breadth and thickness 
and a fourth dimension at right angles to the other three. 
The relativity world of four dimensions we have in reality 
been quite familiar with throughout our whole lives, as is 
signified by the words right-and-left, forward-and-backward, 
up-and-down, and sooner-and-later. The first three dimen- 
sions give the familiar world of objects, the four dimensions 
taken together furnish the world of events. In this four- 
dimensional world a particle occupies not one point, as we 
are accustomed to think of in space of three-dimensions, but 
a series of points representing its positions at successive in- 
stants of time. Its history then is represented by a line, 
called its " world-line." If an observer is admitted, he has 
his own world-line and immediately he imagines that his 
world-line, of all lines, is the most important in the universe. 


He immediately proceeds to divide up the four-dimensional 
entity into a space of three dimensions with time as the 
fourth, and we have the world familiar to us all. But the 
theory of relativity rules out the observer from any con- 
sideration whatever, and there now results a space-time 
nature in which is scarcely recognized space and time. 
There is thus no " shape " to anything, no difference be- 
tween straight and crooked, it is impossible to measure an 
angle, which is in fact the difference in direction between 
two straight lines. This four-dimensional world which was 
invented by Minkowski, naturally has no reality, it does not 
exist and its properties cannot be interpreted by our ordi- 
nary experiences. The fourth dimension at right angles to 
the other three is not even the time plotted as a quantity 
directly, but the product of time into the square root of 
minus one. This quantity, V ~i, is known to every in- 
telligent boy as the symbol of an imaginary quantity. Hence 
it is futile to attempt to assign to this unfamiliar and unreal 
world any of the attributes derived from the common ex- 
perience of life. Many writers in their attempts to popular- 
ize have used such catch-phrases as " time is curved," " there 
is no now" " a phenomenon may be seen before it happens/' 
and so forth and so on. 

" Relegating space and time to their proper source the 
observer Einstein bids us contemplate the residuum of 
what we observe. This residuum is the true world. It is 
shapeless, because we have abstracted shape; yet it is metri- 
cal and has quantitative properties which can be expressed 
in mathematical terms. Clearly we cannot describe this 
true world in terms of familiar things, because the whole 
point of Einstein's theory is that we must subtract the ideas 
which we ourselves have added in order to form familiar 
things. Mathematics is the only language in which the in- 
herent qualities of this unfamiliar world can be described. 
But though it seems unfamiliar, nature is left simpler by this 
purification. A closer unity is perceived in the bases of 
phenomena apparently diverse; and, for example, the effect 
of gravitation on light is clearly foreseen. Further, the laws 
of nature must relate to this four-dimensional residuum and 


the space and time we ourselves introduce cannot be rele- 
vant. This led Einstein to the conclusion that Newton's 
law of gravitation, which refers to one particular separation 
of space and time, cannot be the exact law; and he proposed 
a new law applicable to the four-dimensional world.' 3 1 

Although there is no distinction between straight and 
crooked in the four-dimensional entity, it is however possible 
to join two points or events by a certain unique track which 
plays a part corresponding to the straight line in our familiar 
life. This unique track is called a geodesic, and instead of 
being the shortest distance between two points the im- 
portant property of the straight line in ordinary space 
the geodesic is of the maximum length. We see, therefore, 
what unfamiliar conceptions are involved in the theory of 
relativity, and how radically changed must be the point of 
view that makes the distance between two points a maximum 
and not a minimum ! 

Let us see some of the necessary consequences of the 
principle of relativity in its relation to the propagation of 
light. As already stated, this principle enunciates that the 
velocity of light is independent of the observer, or, stated 
in other words, that no matter what is the velocity of the 
observer the wave-front of an emitted beam of light is al- 
ways a sphere with the observer as the center. According 
to the older views, a definite point of the ether, the source 
of light, was the center of the sphere, but according to the 
newer ideas it is the observer himself who is the center of 
the sphere. The observer is thus in a sense the center of 
the universe, he is at rest and the universe is carried by for 
his inspection. If an observer A is the center of a sphere, 
the wave-front will have advanced after time t so that the 
square of the radius of the sphere will be represented by 

tf + V 2 + Z 2 = C 2 / 2 

or x 1 + f + z 2 - c 2 P = o 

where the observer is the origin of coordinates and c is the 
velocity of light. For any other observer B, the wave-front 

1 Eddington, Contemporary Review, 116, 643, 1919. 


will again be a sphere but with B as its center, and the mathe- 
matical equation will be 

x* + / + z' 2 - c 2 /' 2 = o 

According to the relativity assumption, the sphere observed 
by A is identical with that observed by B. Hence a linear 
transformation between jc, y y z, t and x', y f , z' y t' must trans- 
form either of the above equations into the other. If now 
we put let = w (where I is the imaginary quantity, V J )> 
then the equations reduce to x 1 -f~ T + z 2 + w 2 = o and 
x'~ + y' s + %'* + w '~ = > eac ^ f which represents a 
sphere in four-dimensional space, x, y, z, w form a set of or- 
thogonal coordinates, and x', y', z', w' must represent a 
second set of orthogonal axes in this same space. In space of 
three dimensions we are familiar with such sets of axes. An 
observer at any one place chooses three axes, one north and 
south, the second east and west and the third vertically up 
and down. Another observer at a different place on the 
surface of the earth takes a similar set of axes with reference 
to his directions of north, east and vertical. But what are 
up and down, and north and south to one observer are not 
the same directions that appear to the other observer. A 
rotation of axes and a transfer of the origin will transform 
the set of axes of one observer into those of the other. 
Similarly in the four-dimensional space-time continuum the 
coordinates of A can be transformed into those of B by a 
rotation of axes. But the sphere observed by A being the 
same as that observed by B, or by any other observer, it 
therefore must appear as an objective reality. But A and B 
divide up the space-time continuum differently. Space and 
time depend only on the consciousness of the individual ob- 
server; hence the coordinates that A calls pure space and 
pure time will not be space and time to observer B but will 
be a mixture of the two. It is only the combination of 
space and time, not space and time separately, that repre- 
sents reality independent of the peculiarities of the indi- 
vidual observer. The mathematical expressions by means 
of which x, y, z, w are rotated into x', y', z f y w' are the 


famous transformations of Lorentz which were derived a 
decade before Einstein's work was published. 

Starting with these transformations and assuming that 
light follows a geodesic and not a straight line (which is 
now a meaningless term), Einstein developed his famous 
theory of gravitation. The only constant introduced into 
the discussion is c the velocity of light, the maximum velocity 
of which we have any observational experience. Virtually 
the only assumption made is the one of fundamental im- 
portance, that the velocity of light is identical to all ob- 
servers. This is at best an assumption, and if it can be 
shown by any experiments that this premise is not true, then 
the whole Einstein structure will fall into ruins. By means 
of a little-used and very difficult branch of mathematics, 
a theory was developed which explains gravitation, the 
motion of Mercury and the gravitational attraction of the 
sun on light rays. As the author does not claim to be one 
of the original dozen who could understand Einstein's equa- 
tions he will not introduce any further mathematical formu- 
las. These newer conclusions will herald an advance over 
the old law of gravitation only on the condition that Ein- 
stein's theory represents the observed facts more closely 
than the theory of Newton. 

We may perhaps obtain an inkling of the meaning of 
four-dimensional space by starting out with the properties 
of three-dimensional space, with which from childhood we 
have been familiar. In the world of events all four coor- 
dinates are necessary, for we never observe an event except 
at a certain time, and we are never cognizant of an instant 
of time without special reference to space. In discussing 
the laws of electromagnetic phenomena, it was shown by 
Minkowski that these laws could be represented by assum- 
ing a four-dimensional space-time and that the mathematical 
transformations of Lorentz and Einstein could be described 
by a rotation of a set of axes. In three-dimensional space 
of everyday happenings no one would be impertinent enough 
to say that there is any absolute direction of up and down, 
i.e., the vertical, at any one place on the earth's surface, 
although there are great numbers of people who believe that 


New York City occupies a specially important place on the 
top of the world. The words " sooner or later " are like- 
wise relative terms only and not absolute. We are all fa- 
miliar with the mental process of assuming a set of three 
rectangular axes relating to the particular location of the 
observer on the earth's surface, a set, therefore, of relative 
coordinates, not absolute ones. If we make a section by a 
plane of three-dimensional space, we derive a two-dimen- 
sional space, a plane. Similarly, if a section is made through 
the four-dimensional world of events, a three-dimensional 
space will result. For an observer on the earth, there is one 
particular section only which will give the space of three 
dimensions with which he is familiar, so that the four-dimen- 
sional unity thus breaks up for him into space and ordinary 
time. A section through an observer on Mars would be 
unreal to us and unnatural in aspect. One of the postulates 
regarding four-dimensional space is that it must be entirely 
independent of the observer so that there cannot be any 
real difference between any two directions in an absolute 
sense. For any particular observer, as we have seen, the 
four-dimensional world of events may break up into ordi- 
nary space and time, but such a section would represent a 
particular case and not the most general solution. 

We may draw a number of lines on a flat piece of paper 
or sheet of rubber. These lines intersect in many points. 
By taking the sheet of rubber into our hands and altering 
its shape, the lines on its surface become curves in three- 
dimensional space. A skillful mathematician could repre- 
sent these curves by complicated equations of transforma- 
tions involving x, y and z. By altering the shape of the 
piece of rubber we change the appearance of the lines on its 
surface, but we do not in the least change the number of the 
intersections made by the lines. The transformations ex- 
pressing the intersection of two lines may vary greatly in 
their mathematical forms, but if they are true formulas, they 
cannot alter the actual intersections of the lines themselves. 
A simple case of change of axes of reference, a mathemati- 
cal transformation, is familiar to everyone. While standing 
out in the rain, when no wind is blowing, one holds his um- 


brella vertically over his head if he wishes to keep dry. 
The rain-drops, obeying the law of gravitation, seem to fall 
vertically. If the observer starts to walk, the direction of 
the falling rain appears to have changed, the rain-drops 
come in a slanting direction. Of course, the direction has 
not actually changed, but the effect is just the same as if 
there had been a real change in the direction of gravity. 
Some obstinate person will say that he knows that gravity 
is acting in a vertical direction and that his motion has noth- 
ing at all to do with gravity. Under this assumption he 
persists in holding the handle of his umbrella vertical, for 
he has learned that when standing at rest he keeps dry by 
so doing. If he still insists in his foolishness when going at 
high speed in a swift-moving, open automobile, he will lose 
his umbrella and get dripping wet into the bargain. 

In four-dimensional space the " world-lines " intersect in 
a series of events. For the description of the world-line we 
can choose any set of four-dimensional axes we wish. The 
event takes place, however, absolutely independently of 
any assumption of reference axes. If we change the axes, 
we at the same time change the coordinates with respect to 
these axes, but the interesting thing is the event and not 
the particular set of reference axes assumed. The world of 
nature is made known to us only through a series of obser- 
vations or series of events. We learn, in fact, only through 
a series of coincidences. To represent the laws of nature 
our observations must be true no matter what selection we 
make of reference axes. In other words, the mathematical 
expression of the laws of nature must be such that their form 
does not change if we make a transformation of axes. 

It is a curious fact that we know very little of the mecha- 
nism by which gravitation works. Is it propagated with the 
velocity of light, or does it act instantaneously at infinite 
speed? At a place on the earth's surface, gravity is the 
resultant of two forces, one the attraction of the earth, the 
other due to the centrifugal force of the earth's rotation. 
The former is spoken of generally as a " natural " force, 
the latter as an " artificial " one. Centrifugal force may be 
altered by taking different locations on the surface of the 


earth or may be made to vanish entirely by going to the 
north or south poles. For the purpose of simplicity in the 
mathematical treatment, centrifugal force is separated from 
gravity. We have no direct sensation for either force sepa- 
rately, and in fact there is no physical basis for the separa- 
tion. So why not consider the " natural " and the " arti- 
ficial " force on the same basis? The generalization of this 
notion led to Einstein's principle of equivalence that a 
gravitational field of force is precisely equivalent to an 
artificial field of force, so that it is impossible by any con- 
ceivable experiment to distinguish between them. Force, 
therefore, is relative and not absolute. 

Guided by the two principles of relativity and equivalence, 
Einstein was led to assume that all gravitational fields of 
force must be illusions. He himself admits that he was 
brought to his new point of view by discussing the sensa- 
tions of falling with a man who had just tumbled from a 
high building. No distinct consciousness of falling was 
actually experienced by the man. The simplest method 
after all for the consideration of gravitation may actually 
be the point of view of the falling man who was experiencing 
no peculiar sensations of his sudden flight. This was 
Einstein's method. The earth was surely rising to meet 
the man, and if he had not known of the sudden stoppage 
of his flight that was in store for him, he might have quietly 
theorized regarding the relativity of motion. Such an un- 
constrained body (as the falling man) if left to itself takes 
what we shall call a " straight " line, but which is actually 
a geodesic. The world-line is undeflected until the particle 
of matter comes into the vicinity of another particle. The 
world-line of each particle is bent in towards each other or 
deformed. A change of direction of the particle means that 
a field of force has been entered. Such a deformation can 
be brought into the equations by means of mathematical 

Gravitation therefore needs no special treatment different 
from that of any other force. In developing the mathe- 
matical analysis it was not possible to follow the geometry 
of the flat space of Euclid where the straight line is the 


shortest distance between two points. The mathematician 
will assert that there is no reason to assume Euclidian 
geometry unless observation demands it. In the process of 
the difficult mathematical development, Einstein was at 
liberty to choose between several possibilities, and the deci- 
sion reached seemed to give the simplest solution. In the 
final analysis his law of gravitation must produce the same 
effect as Newton's law for conditions of velocities small 
compared with that of light. 

According to Einstein's views, gravitation is simply a 
geometrical deformation of unrestrained bodies, and hence 
we may regard the gravitational field as influencing or even 
determining the laws of space-time derived from measure- 
ments. The track of a ray of light therefore is deformed 
by the gravitational field. According to Eddington 1 " a 
ray of light passing near a heavy particle will be bent, 
firstly, owing to the non-Euclidian character of the com- 
bination of time with space. This bending is equivalent to 
that due to the Newtonian gravitation, and may be cal- 
culated in the ordinary way on the assumption that light 
has weight like a material body. Secondly, it will be bent 
owing to the non-Euclidian character of space alone, and 
this curvature is additional to that predicted by Newton's 
law. If then we can observe the amount of curvature of a 
ray of light, we can make a crucial test of whether Einstein's 
or Newton's theory is obeyed. 

" This separation of the attraction into two parts is useful 
in a comparison of the new theory with the old, but from 
the point of view of relativity it is artificial. Our view is 
that light is bent in just the same manner as the track of a 
material particle moving with the same velocity would be 
bent. Both causes of bending may be ascribed either to 
weight or to non-Euclidian space-time, according to the 
nomenclature preferred. The only difference between the 
predictions of the old and new theories is that in one case 
the weight is calculated according to Newton's law of gravi- 
tation, in the other case according to Einstein's." 

It has been repeatedly urged by many writers that New- 

1 Space y Time and Gravitation, p. 106. 


ton's law is simpler than Einstein's. The former is indeed 
more familiar, but that does not necessarily mean that it is 
simpler. This depends on the point of view. The principle of 
relativity introduced into scientific thought, first destroyed 
the notion that space and time were absolute in character 
or objective in existence. This led to a consideration of the 
four-dimensional world of events and to the supposition that 
gravitational force is an illusion. To explain this " force " 
as an inherent curvature in space represents a further 
revolution in scientific thought. Gravitation is not the only 
" force " familiar to the physicist. Are these also illusions 
and must not the Einstein theory be generalized to include 
all of them? Such a generalization has been proposed by 
Weyl who of necessity was forced to introduce new curva- 
tures in the four-dimensional space to represent the forces 
involved ; in fact, the effects predicted by Weyl agree so per- 
fectly with electromagnetic theory that no testing of the 
theory by experiment is possible. The only scientific test 
to which all theories and laws must submit is the funda- 
mental one of whether they represent the observed facts. 
Einstein has not attempted an explanation of gravitation 
other than to assume that it is propagated with the velocity 
of light; he has only been occupied with the deduction of its 

In the foregoing chapter is given an inkling of the man- 
ner in which the philosophic trend of thought has been modi- 
fied by the theory of relativity; in the following pages some 
few of the scientific consequences will be discussed. 



IF the path of a ray of light is identical with that of a 
material particle when passing close to a body of heavy 
mass like the sun, it is easy to see that the path will not 
be straight but will be a hyperbola. To calculate the amount 
of the deflection from the straight line course one might 
apply Newton's law; but according to Einstein's theory, this 
old and well established law is applicable only under the con- 
dition of small velocities. For an object traveling with the 
velocity of light, the Einstein formula gives a deflection 
twice that deduced from Newton's law. 

Other important consequences follow as a direct result 
of the Einstein theory. According to the inverse-square 
law of gravitation, an isolated planet will describe about 
the sun an ellipse the direction of whose axis is fixed in space, 
but according to the Einstein law the path will be a spiral 
and not an ellipse. The direction of the major axis will 
therefore rotate in space instead of being stationary. The 
amount of rotation, calculated from the Einstein theory and 
derived without the introduction of any additional constants 
into the equations, is 43 " per century for the planet 

A further consequence is found when comparing the time 
of vibration of an atom acting under the strong gravita- 
tional pull of the sun with that of a similar atom in a ter- 
restrial laboratory. The atom in its vibrations always be- 
haves very much like the pendulum of an ideal clock. Since 
the gravitational field of force is similar in its properties to 
that of a centrifugal field, it is comparatively easy to calcu- 
late the ef